diff --git a/about/index.html b/about/index.html index c5a8251..387afec 100644 --- a/about/index.html +++ b/about/index.html @@ -87,7 +87,7 @@ diff --git a/archives/2023/08/index.html b/archives/2023/08/index.html index 0fa4d54..101a7f3 100644 --- a/archives/2023/08/index.html +++ b/archives/2023/08/index.html @@ -84,7 +84,7 @@

尚 Shawn's Personal Page

diff --git a/archives/2023/09/index.html b/archives/2023/09/index.html index b855ade..f4a4ca0 100644 --- a/archives/2023/09/index.html +++ b/archives/2023/09/index.html @@ -84,7 +84,7 @@

尚 Shawn's Personal Page

diff --git a/archives/2023/11/index.html b/archives/2023/11/index.html index 0ff2963..0c58cb6 100644 --- a/archives/2023/11/index.html +++ b/archives/2023/11/index.html @@ -84,7 +84,7 @@

尚 Shawn's Personal Page

diff --git a/archives/2023/12/index.html b/archives/2023/12/index.html index 90b59bd..829bc76 100644 --- a/archives/2023/12/index.html +++ b/archives/2023/12/index.html @@ -84,7 +84,7 @@

尚 Shawn's Personal Page

diff --git a/archives/2023/index.html b/archives/2023/index.html index 3f07b55..77035eb 100644 --- a/archives/2023/index.html +++ b/archives/2023/index.html @@ -84,7 +84,7 @@

尚 Shawn's Personal Page

diff --git a/archives/2023/page/2/index.html b/archives/2023/page/2/index.html index 9764bdd..bd1edc2 100644 --- a/archives/2023/page/2/index.html +++ b/archives/2023/page/2/index.html @@ -84,7 +84,7 @@

尚 Shawn's Personal Page

diff --git a/archives/2024/01/index.html b/archives/2024/01/index.html index e69d26f..735909c 100644 --- a/archives/2024/01/index.html +++ b/archives/2024/01/index.html @@ -84,7 +84,7 @@

尚 Shawn's Personal Page

diff --git a/archives/2024/index.html b/archives/2024/index.html index 530d5d9..8a74a58 100644 --- a/archives/2024/index.html +++ b/archives/2024/index.html @@ -84,7 +84,7 @@

尚 Shawn's Personal Page

diff --git a/archives/index.html b/archives/index.html index f471d25..d168c72 100644 --- a/archives/index.html +++ b/archives/index.html @@ -84,7 +84,7 @@

尚 Shawn's Personal Page

diff --git a/archives/page/2/index.html b/archives/page/2/index.html index a70ff95..31e5f5a 100644 --- a/archives/page/2/index.html +++ b/archives/page/2/index.html @@ -84,7 +84,7 @@

尚 Shawn's Personal Page

diff --git a/archives/page/3/index.html b/archives/page/3/index.html index 4483f8e..d5274f0 100644 --- a/archives/page/3/index.html +++ b/archives/page/3/index.html @@ -84,7 +84,7 @@

尚 Shawn's Personal Page

diff --git a/baidusitemap.xml b/baidusitemap.xml index 28c6145..bbe8af7 100644 --- a/baidusitemap.xml +++ b/baidusitemap.xml @@ -4,168 +4,84 @@ https://captaincccp.github.io/%E3%80%90WORK-development%20record%E3%80%91EQCat/ 2024-10-14 - - https://captaincccp.github.io/blog/%E3%80%90WORK-development%20record%E3%80%91EQCat.html - 2024-10-14 - https://captaincccp.github.io/%E3%80%90WORK-intern%20experience%E3%80%91BeijingAcademyOfScienceAndTechnology%20-%20%E5%89%AF%E6%9C%AC/ 2024-10-14 - - https://captaincccp.github.io/blog/%E3%80%90WORK-intern%20experience%E3%80%91BeijingAcademyOfScienceAndTechnology%20-%20%E5%89%AF%E6%9C%AC.html - 2024-10-14 - https://captaincccp.github.io/%E3%80%90AREA-introduction%E3%80%91LLM/ 2024-03-06 - - https://captaincccp.github.io/blog/%E3%80%90AREA-introduction%E3%80%91LLM.html - 2024-03-06 - https://captaincccp.github.io/%E3%80%90BOOK-reading%20note%E3%80%91The%20Three%20Penny%20Opera/ 2024-03-06 - - https://captaincccp.github.io/blog/%E3%80%90BOOK-reading%20note%E3%80%91The%20Three%20Penny%20Opera.html - 2024-03-06 - https://captaincccp.github.io/%E3%80%90MATH-notes%E3%80%91N=NP%20problem/ 2023-12-22 - - https://captaincccp.github.io/blog/%E3%80%90MATH-notes%E3%80%91N=NP%20problem.html - 2023-12-22 - https://captaincccp.github.io/%E3%80%90OTHER-novel%E3%80%91AI%20novel%20competition/ 2023-12-22 - - https://captaincccp.github.io/blog/%E3%80%90OTHER-novel%E3%80%91AI%20novel%20competition.html - 2023-12-22 - https://captaincccp.github.io/%E3%80%90DOCS-notes%E3%80%91how%20to%20build%20a%20NAS/ 2023-12-19 - - https://captaincccp.github.io/blog/%E3%80%90DOCS-notes%E3%80%91how%20to%20build%20a%20NAS.html - 2023-12-19 - https://captaincccp.github.io/%E3%80%90PAPER-introduction%E3%80%91VideoGPT/ 2023-12-06 - - https://captaincccp.github.io/blog/%E3%80%90PAPER-introduction%E3%80%91VideoGPT.html - 2023-12-06 - https://captaincccp.github.io/%E3%80%90PHIL-guide%E3%80%91What%20Causality%20tells%20in%20real%20life/ 2023-12-05 - - https://captaincccp.github.io/blog/%E3%80%90PHIL-guide%E3%80%91What%20Causality%20tells%20in%20real%20life.html - 2023-12-05 - https://captaincccp.github.io/%E3%80%90DOCS-notes%E3%80%91ubuntu/ 2023-12-05 - - https://captaincccp.github.io/blog/%E3%80%90DOCS-notes%E3%80%91ubuntu.html - 2023-12-05 - https://captaincccp.github.io/%E3%80%90TRANS-work%E3%80%91Marsilio%20Ficino/ 2023-12-05 - - https://captaincccp.github.io/blog/%E3%80%90TRANS-work%E3%80%91Marsilio%20Ficino.html - 2023-12-05 - https://captaincccp.github.io/%E3%80%90PAPER-summary%E3%80%91Papers%20read%2023/ 2023-12-05 - - https://captaincccp.github.io/blog/%E3%80%90PAPER-summary%E3%80%91Papers%20read%2023.html - 2023-12-05 - https://captaincccp.github.io/%E3%80%90AREA-introduction%E3%80%91Generative%20models/ 2023-12-05 - - https://captaincccp.github.io/blog/%E3%80%90AREA-introduction%E3%80%91Generative%20models.html - 2023-12-05 - https://captaincccp.github.io/%E3%80%90BOOK-reading%20note%E3%80%91Shakespeare%20Speaks/ 2023-12-05 - - https://captaincccp.github.io/blog/%E3%80%90BOOK-reading%20note%E3%80%91Shakespeare%20Speaks.html - 2023-12-05 - https://captaincccp.github.io/%E3%80%90BOOK-reading%20note%E3%80%91The%20Hairy%20Ape/ 2023-12-05 - - https://captaincccp.github.io/blog/%E3%80%90BOOK-reading%20note%E3%80%91The%20Hairy%20Ape.html - 2023-12-05 - https://captaincccp.github.io/%E3%80%90PAPER-introduction%E3%80%91VQ-VAE/ 2023-12-05 - - https://captaincccp.github.io/blog/%E3%80%90PAPER-introduction%E3%80%91VQ-VAE.html - 2023-12-05 - https://captaincccp.github.io/%E3%80%90MATH-notes%E3%80%91Differential%20Equation/ 2023-12-05 - - https://captaincccp.github.io/blog/%E3%80%90MATH-notes%E3%80%91Differential%20Equation.html - 2023-12-05 - https://captaincccp.github.io/%E3%80%90AREA-introduction%E3%80%91time%20series/ 2023-12-05 - - https://captaincccp.github.io/blog/%E3%80%90AREA-introduction%E3%80%91time%20series.html - 2023-12-05 - https://captaincccp.github.io/%E3%80%90PAPER-introduction%E3%80%91DDPM/ 2023-12-05 - - https://captaincccp.github.io/blog/%E3%80%90PAPER-introduction%E3%80%91DDPM.html - 2023-12-05 - https://captaincccp.github.io/%E3%80%90PAPER-introduction%E3%80%91CausalVAE/ 2023-12-05 - - https://captaincccp.github.io/blog/%E3%80%90PAPER-introduction%E3%80%91CausalVAE.html - 2023-12-05 - https://captaincccp.github.io/%E3%80%90AREA-introduction%E3%80%91Causal%20Science/ 2023-12-05 - - https://captaincccp.github.io/blog/%E3%80%90AREA-introduction%E3%80%91Causal%20Science.html - 2023-12-05 - \ No newline at end of file diff --git a/blog/index.html b/blog/index.html index b1034e0..31fcbe0 100644 --- a/blog/index.html +++ b/blog/index.html @@ -84,7 +84,7 @@

尚 Shawn's Personal Page

diff --git a/blog/page/2/index.html b/blog/page/2/index.html index 2f1f09e..4226d43 100644 --- a/blog/page/2/index.html +++ b/blog/page/2/index.html @@ -84,7 +84,7 @@

尚 Shawn's Personal Page

diff --git a/blog/page/3/index.html b/blog/page/3/index.html index bfc5926..6cc54c2 100644 --- a/blog/page/3/index.html +++ b/blog/page/3/index.html @@ -84,7 +84,7 @@

尚 Shawn's Personal Page

diff --git "a/blog/\343\200\220AREA-introduction\343\200\221Causal Science.html" "b/blog/\343\200\220AREA-introduction\343\200\221Causal Science.html" deleted file mode 100644 index 1a75850..0000000 --- "a/blog/\343\200\220AREA-introduction\343\200\221Causal Science.html" +++ /dev/null @@ -1,304 +0,0 @@ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -【AREA-introduction】Causal Science | 尚 Shawn's Personal Page - - - - - - - - - - - - - - -
- -
-
-
- - - - - -
- - - - - - - - -
- - - - - -
- -
- - - - - -
- -

【AREA-introduction】Causal Science -

- - - -
- - - -
-

causal AI learning (building)

- -

Background knowledges

基于约束的方法

Graph

    -
  • 3 kinds:

    -
  • -
  • chain X->Z->Y

    -
  • -
  • fork X<-Z->Y
    Z是X和Y的共因,如果X发生改变,Y很有可能也应该改变
    when the middle Z is given,X and Y won’t be affected anymore.

    -
  • -
  • collider x->Z<-Y
    在没有given Z时,X和Y是独立的
    when the middle Z is given,X/Y would need to change to make sure Z won’t change while Y/X is changing

    -
  • -
  • Path can’t be blocked by S(node set).

    -
  • -
  • blocked:

    -
      -
    • given Z when chain/fork
    • -
    • not given Z(and its son nodes) when collider
    • -
    -
  • -
  • D-Seperation:
    when every path between X and Y is blocked by Z

    -
  • -
  • Markovian assumption

    -
  • -
  • Faithfulness assumption
    由于以上两个假设我们可以通过given数据推出对应的Markov等价类

    -
  • -
-

典型算法

IC、PC、FCI、RFCI、GFCI
基本思想是1.通过数据找到条件独立性列表(cond. independences) 2.通过条件独立性列表得到DAGs有向无环图

-

Markov Equivalence Class马尔科夫等价类

    -
  • 可以由同一组d-seperation得到的图s
  • -
  • 马尔科夫等价类中的图有相同的骨架skeleton和V-结构
    比如chain和fork的3中图就有相同的V结构
  • -
-

基于分数的方法 score-based

Causal inference

-
- - - -
- - - - - -
-
- - - - -
- - 0% -
- - - - - - - - - - - - - - - - - - - - - - - - - diff --git "a/blog/\343\200\220AREA-introduction\343\200\221Generative models.html" "b/blog/\343\200\220AREA-introduction\343\200\221Generative models.html" deleted file mode 100644 index 00dc50a..0000000 --- "a/blog/\343\200\220AREA-introduction\343\200\221Generative models.html" +++ /dev/null @@ -1,281 +0,0 @@ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -【AREA-introduction】Generative models | 尚 Shawn's Personal Page - - - - - - - - - - - - - - -
- -
-
-
- - - - - -
- - - - - - - - -
- - - - - -
- -
- - - - - -
- -

【AREA-introduction】Generative models -

- - - -
- - - -
-
Introduction of common generative models. (building)
-
- -

general

Most of the datas such as audios,texts,images in our daily life have a lot of redundant informations.

-

VAEs

AE

Its biggest difference with normal networks is that there are much fewer nodes at its hidden layer,
which is like a bottleneck,that forced the network to learn about the specific representation of the input.
The encoder would be trained based on the difference between the reconstructed output and input.

-

But,due to the overfitting of the encoder,we could not possibly use the decoder as a generator.

-

VAE

The main task of VAE’s encoder is to learn how to represent 2 fractors mean and variance of the origin distribution with
a new distribution at hidden layer level,that is,P(X|Z)

-

By

-

VQ-VAE

Flow

GAN

DDPM

-
- - - -
- - - - - -
-
- - - - -
- - 0% -
- - - - - - - - - - - - - - - - - - - - - - - - - diff --git "a/blog/\343\200\220AREA-introduction\343\200\221LLM.html" "b/blog/\343\200\220AREA-introduction\343\200\221LLM.html" deleted file mode 100644 index fdeb368..0000000 --- "a/blog/\343\200\220AREA-introduction\343\200\221LLM.html" +++ /dev/null @@ -1,1002 +0,0 @@ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -【AREA-introduction】LLM | 尚 Shawn's Personal Page - - - - - - - - - - - - - - -
- -
-
-
- - - - - -
- - - - - - - - -
- - - - - -
- -
- - - - - -
- -

【AREA-introduction】LLM -

- - - -
- - - -
-
Brief survey of LLM for project. 
-
- - -

模型能力排行

综合能力:
综合能力
分类能力:
信息抽取能力:
阅读理解能力:
数据分析能力:

-

quick link

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
大模型机构类别链接价格contextwindow发布时间备注
GPT-4 Turboopenai商用https://openai.com/productInput: $0.01 / 1K tokens Output: $0.03 / 1K tokens”128K
GPT-4openai商用https://openai.com/product“Input: $0.03 / 1K tokens Output: $0.06 / 1K tokens”~25K
GPT-4-32Kopenai商用https://openai.com/product“Input: $0.06 / 1K tokens Output: $0.12 / 1K tokens”32K
GPT-3.5-turbo-1106openai商用https://openai.com/product“ $0.0010 / 1K tokens $0.0020 / 1K tokens”16K
GPT-3.5-turbo-instructopenai商用https://openai.com/product“ $0.0015 / 1K tokens $0.0020 / 1K tokens”4K
Claude InstantAnthropic商用https://www.anthropic.com/product“$0.80/million tokens $2.40/million tokens”100K
Claude 2.0Anthropic商用“$8.00/million tokens $24.00/million tokens”100K
Claude 2.1Anthropic商用“$8.00/million tokens $24.00/million tokens”200K
Gemini Pro谷歌商用https://deepmind.google/technologies/gemini/#introduction“free:60 chat / min In: $0.00025 / 1K characters ou:$0.0005 / 1K char”不得超过 4MB
PaLM谷歌商用
Bard谷歌商用
文心一言4.0(ERNIE-Bot 4.0)百度商用“¥0.12 / 1K tokens ¥0.12 / 1K tokens”2023.6价格网页
文心一言3.5 8K(ERNIE-Bot-8K)百度商用“¥0.024 / 1K tokens ¥0.048 / 1K tokens”大模型平台
文心一言3.5(ERNIE-Bot)百度商用“¥0.012 / 1K tokens ¥0.012 / 1K tokens”chat页面
文心一言EB-turbo-AppBuilder专用版百度商用“¥0.008 / 1K tokens ¥0.008 / 1K tokens”
讯飞星火v3.0科大讯飞商用https://xinghuo.xfyun.cn/sparkapi ~¥0.03 / 1K tokens免费试用很多
讯飞星火v1.5科大讯飞商用~¥0.015 / 1K tokens
腾讯混元标准版腾讯商用https://cloud.tencent.com/product/hunyuan¥0.01 / 1K
腾讯混元高级版腾讯商用https://cloud.tencent.com/product/hunyuan¥0.1 / 1K
“天工 SkyChat-MegaVerse”昆仑万维商用https://model-platform.tiangong.cn/?channel=csearch_enter&utm_medium=csearch¥0.005 / 1K做一个agent活动针对长文本
AGI Sky-Saas-Search昆仑万维商用¥0.18 / 1K针对搜索
AGI Sky-Saas-Writing昆仑万维商用需要咨询针对写作
Skywork-13B-Base昆仑万维开源https://model-platform.tiangong.cn/开源4K
Baichuan2-Turbo百川智能商用https://www.baichuan-ai.com/home0.008元/千tokens
Baichuan2-Turbo-192k百川智能商用0.016元/千tokens192K
Baichuan2-53B百川智能商用“8点前0.01元/千tokens 其余时间0.02”
Baichuan2-7B百川智能开源开源4K
Baichuan2-13B百川智能开源开源4K
baichuan-7B百川智能开源开源4K
baichuan-13B-chat百川智能开源https://github.com/baichuan-inc/Baichuan-13B开源4K
华为盘古华为
ChatGLM-6B智谱AI开源开源github
智谱清源智谱AI商用
360智脑奇虎360商用https://ai.360.cn/
通义千问阿里巴巴商用https://tongyi.aliyun.com
通义千问Qwen-1.8B阿里巴巴开源开源32Kgithub
通义千问Qwen-7B阿里巴巴开源开源32K
通义千问Qwen-14B阿里巴巴开源开源8K
通义千问Qwen-72B阿里巴巴开源开源32K
minimaxminimax商用https://api.minimax.chat
tigerbot-7b-chat虎博科技商用/开源https://www.tigerbot.com/0.018元 / 千tokensapi界面
tigerbot-13b-chat虎博科技商用https://tigerbot.com/api-reference/chat0.02元 / 千tokens
tigerbot-70b-chat虎博科技0.04元 / 千tokens
chatglm-6b清华大学&智谱AI开源https://github.com/THUDM/ChatGLM-6B 开源
LLaMA-2META开源开源
LlaMA-13BMETA开源开源
Kimi Chat月之暗面商用https://www.yjpoo.com/site/1809.htmlchat
belle-llama-7b-2m链家科技开源https://github.com/LianjiaTech/BELLE开源
BELLE-on-Open-Datasets链家科技开源https://github.com/LianjiaTech/BELLE开源
belle-llama-13b-2m链家科技开源https://github.com/LianjiaTech/BELLE开源
belle-llama-13b-ext链家科技开源https://github.com/LianjiaTech/BELLE开源
Ziya-LLaMA-13B-v1IDEA研究院开源https://mp.weixin.qq.com/s/IeXgq8blGoeVbpIlAUCAjA开源
phoenix-inst-chat-7b港中文开源https://github.com/FreedomIntelligence/LLMZoo开源
linly-chatflow-13b深圳大学开源https://github.com/CVI-SZU/Linly开源
MOSS-003-SFT复旦大学开源https://github.com/OpenLMLab/MOSS开源
AquilaChat-7B智源研究院开源https://github.com/FlagAI-Open/FlagAI/blob/master/examples/Aquila/README.md开源
tulu-30ballenai开源https://github.com/allenai/open-instruct开源
-

info

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
参数显存(推理)显存(微调)备注
70B模型~48GB~62GB
14B模型~13GB~19GB
7B模型~8GB~12GB
2B模型~3GB~6GB
-

上下文窗口:

192K大约是35万汉字,100K大约8万汉字,32K大约2.5万字

-

database

    -
  • MySQL
  • -
  • MongoDB
  • -
  • Redis
  • -
  • Chroma is the open-source embedding database.
  • -
  • Weaviate is an open source, AI-native vector database
  • -
  • Milvus向量数据库
  • -
  • pinecone
  • -
-

agent

- -
- - - -
- - - - - -
-
- - - - -
- - 0% -
- - - - - - - - - - - - - - - - - - - - - - - - - diff --git "a/blog/\343\200\220AREA-introduction\343\200\221time series.html" "b/blog/\343\200\220AREA-introduction\343\200\221time series.html" deleted file mode 100644 index f8b8c2b..0000000 --- "a/blog/\343\200\220AREA-introduction\343\200\221time series.html" +++ /dev/null @@ -1,340 +0,0 @@ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -【AREA-introduction】time series | 尚 Shawn's Personal Page - - - - - - - - - - - - - - -
- -
-
-
- - - - - -
- - - - - - - - -
- - - - - -
- -
- - - - - -
- -

【AREA-introduction】time series -

- - - -
- - - -
-
Time series tasks, RNN LSTM and Transformers. (building)
-
- - -

catalog

    -
  1. Introduction

    -
      -
    • time series task
    • -
    -
  2. -
  3. Traditional methods

    -
      -
    • prophet
    • -
    • ARIMA
    • -
    -
  4. -
  5. DeepLearning methods

    -
      -
    • RNN and LSTM
    • -
    • Transformers
        -
      • Vanilla Transformer
      • -
      • Informer
      • -
      • AutoFormer
      • -
      • CrossFormer
      • -
      • PatchTST
      • -
      • FEDFormer
      • -
      • TSAT
      • -
      • iTransformerop
      • -
      -
    • -
    -
  6. -
  7. Spatial and Temporal

    -
  8. -
-

Introduction

time series task

A reliable and efficient representation of multivariate time series is crucial in various downstream machine learning tasks.
In multivariate time series forecasting, each variable depends on its historical values and there are inter-dependencies among variables as well.
Models have to be designed to capture both intra and inter relationships among the time series.

-

Traditional methods

overall

These method like ARIMA and ETS for example,have statistical interpretability while lack adaptability and are limited when trying to capture non-linear relationships.

-

prophet

ARIMA (autoregressive integrated moving average)

ETS (exponential smoothing)

shapelets

DeepLearning methods

RNN and LSTM:

RNN (Recurrent Neural Networks),is designed to process sequential data.
LSTM (Long Short-Term Memory) is designed to capture both the short and longterm patterns within sequential inputs and alleviate gradient vanishing and exploding problems in RNN models

-

Transformers:

    -
  • Vanilla Transformer

    -
  • -
  • Informer (AAAI 2020)

    -
  • -
  • Autoformer (NeurIPS 2021)

    -
  • -
  • CrossFormer

    -
  • -
  • PatchTST

    -
  • -
  • FEDformer (2022)

    -
  • -
  • TSAT (2022)
    this method is transformer with GNN

    -
  • -
  • iTransformer
    iTransformer is a new method came out in 2023 which shows SOTA performance now.

    -
  • -
-

Basicly this model adopted some slight changes based on vanilla Transformer to allow:
1.project the origin series into “Variate Token”,which means instead of including 1 step of all variates,the token now include all step of 1 variate.

-

Spatial and Temporal

    -
  • Traffic Transformer (GIS 2020)
  • -
-

Future

    -
  • Inductive Biases for Time Series Transformers
  • -
  • Transformers and GNN for Time Series
  • -
  • Pre-trained Transformers for Time Series
  • -
  • Transformers with NAS for Time Series
  • -
  • -
-

cite

https://zhuanlan.zhihu.com/p/484044005
https://blog.csdn.net/weixin_36488653/category_11689003.html

- -
- - - -
- - - - - -
-
- - - - -
- - 0% -
- - - - - - - - - - - - - - - - - - - - - - - - - diff --git "a/blog/\343\200\220BOOK-reading note\343\200\221Shakespeare Speaks.html" "b/blog/\343\200\220BOOK-reading note\343\200\221Shakespeare Speaks.html" deleted file mode 100644 index 9718a93..0000000 --- "a/blog/\343\200\220BOOK-reading note\343\200\221Shakespeare Speaks.html" +++ /dev/null @@ -1,276 +0,0 @@ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -【BOOK-reading note】Shakespeare Speaks | 尚 Shawn's Personal Page - - - - - - - - - - - - - - -
- -
-
-
- - - - - -
- - - - - - - - -
- - - - - -
- -
- - - - - -
- -

【BOOK-reading note】Shakespeare Speaks -

- - - -
- - - -
-

Notes for “Shakespeare speaks” series of video.

- - -

What’s done is done

The green-eyed monster

Greek to me

Though this be madness, yet there is method in’t

Wear my heart upon my sleeve

All that glisters is not gold

in a pickle

I must be cruel, only to be kind

As dead as a doornail

A pound of flesh

Give no words but mum

I’ll send him packing

Strange bedfellows

Spotless reputation

Forever and a day

Not budge an inch

The words mine oyster

A fools paradise

Wild-goose chase

A tower of strength

-
- - - -
- - - - - -
-
- - - - -
- - 0% -
- - - - - - - - - - - - - - - - - - - - - - - - - diff --git "a/blog/\343\200\220BOOK-reading note\343\200\221The Hairy Ape.html" "b/blog/\343\200\220BOOK-reading note\343\200\221The Hairy Ape.html" deleted file mode 100644 index 2d2e57d..0000000 --- "a/blog/\343\200\220BOOK-reading note\343\200\221The Hairy Ape.html" +++ /dev/null @@ -1,274 +0,0 @@ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -【BOOK-reading note】The Hairy Ape | 尚 Shawn's Personal Page - - - - - - - - - - - - - - -
- -
-
-
- - - - - -
- - - - - - - - -
- - - - - -
- -
- - - - - -
- -

【BOOK-reading note】The Hairy Ape -

- - - -
- - - -
-

Notes after reading the great play The Hairy Ape by Eugene O’Neill.

- - -
- - - -
- - - - - -
-
- - - - -
- - 0% -
- - - - - - - - - - - - - - - - - - - - - - - - - diff --git "a/blog/\343\200\220BOOK-reading note\343\200\221The Three Penny Opera.html" "b/blog/\343\200\220BOOK-reading note\343\200\221The Three Penny Opera.html" deleted file mode 100644 index 4c2ded3..0000000 --- "a/blog/\343\200\220BOOK-reading note\343\200\221The Three Penny Opera.html" +++ /dev/null @@ -1,280 +0,0 @@ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -【BOOK-reading note】The Three Penny Opera | 尚 Shawn's Personal Page - - - - - - - - - - - - - - -
- -
-
-
- - - - - -
- - - - - - - - -
- - - - - -
- -
- - - - - -
- -

【BOOK-reading note】The Three Penny Opera -

- - - -
- - - -
-

Notes after reading part of “The Three Penny Opera” by Bertolt Brecht.

- - -

song of human weakness

origin text

“It’s called the song of human wekness. Have a listen. You might learn something.

-

A man lives from his head, and tries to use his wits.
But take a look inside his head, and all you’ll find is nits.
In a bleak existence, man is never smart enough.
He will never go the distance, if you call his bluff.

-

You make yourself a plan, you crawl right to the top.
Then, steal yourself another plan, but watch the whole thing flop.
In a cruel existence, man is never bad enough.
All of his mad insistence leaves him in the buff.

-

So, try to chase your luck, but don’t chase luck too fast.
When every fucker’s chasing luck, the luck can never last.
In a poor existence, man is never pure enough.
All of his grand insistence, just a load of guff.”

- -
- - - -
- - - - - -
-
- - - - -
- - 0% -
- - - - - - - - - - - - - - - - - - - - - - - - - diff --git "a/blog/\343\200\220DOCS-notes\343\200\221how to build a NAS.html" "b/blog/\343\200\220DOCS-notes\343\200\221how to build a NAS.html" deleted file mode 100644 index 57d0378..0000000 --- "a/blog/\343\200\220DOCS-notes\343\200\221how to build a NAS.html" +++ /dev/null @@ -1,300 +0,0 @@ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -【DOCS-notes】how to build a NAS | 尚 Shawn's Personal Page - - - - - - - - - - - - - - -
- -
-
-
- - - - - -
- - - - - - - - -
- - - - - -
- -
- - - - - -
- -

【DOCS-notes】how to build a NAS -

- - - -
- - - -
-

What is NAS and how to build one by yourself.

- - -

What and why

NAS(Network Attached Storage)

So basically it is a small server with giant storage that you can access from anywhere, only if you can access the Internet.

-
    -
  • You will have an iPhone or macbook with 60 times cheaper storage.
  • -
  • All the movies, datasets, p…ictures, chatting records, blablabla, will find their place at the disks.
  • -
  • You can deploy websites, dockers, which will be a lot of fun.
  • -
-

How

Hardware

So there are always these 2 choices, buying prebuild NAS or DIY by yourself.

-
    -
  1. if you don’t want to spend any extra attention or time on building or maintaining, just go buy a NAS system from any brand you like, for example Synology.
    Pros:
  2. -
-
    -
  • convenient. You just need to buy it, and buy some disks, put them together and follow the tutorial, done.
  • -
  • after sales services. Usually, you will have a expert who can solve any problems you met and provides some advices.
  • -
  • good looking. prebuild
    Cons:
  • -
  • expensive. Normally it takes a lot more than DIY.
  • -
  • a little bit stiff. You can’t usually upgrade it any further. For example, you can’t simply buy a better network card and attach it. And the operating system is probably locked to the brand’s own system.
  • -
  • privacy. I know it is promised, but you never know.
  • -
-
    -
  1. buy components yourself
    Pros:
  2. -
-
    -
  • cheap.
  • -
- -
- - - -
- - - - - -
-
- - - - -
- - 0% -
- - - - - - - - - - - - - - - - - - - - - - - - - diff --git "a/blog/\343\200\220DOCS-notes\343\200\221ubuntu.html" "b/blog/\343\200\220DOCS-notes\343\200\221ubuntu.html" deleted file mode 100644 index ff24653..0000000 --- "a/blog/\343\200\220DOCS-notes\343\200\221ubuntu.html" +++ /dev/null @@ -1,297 +0,0 @@ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -【DOCS-notes】ubuntu | 尚 Shawn's Personal Page - - - - - - - - - - - - - - -
- -
-
-
- - - - - -
- - - - - - - - -
- - - - - -
- -
- - - - - -
- -

【DOCS-notes】ubuntu -

- - - -
- - - -
-

linux terminal

- -

conda

conda remove –name test –all

-

install

tar.gz paks

-
    -
  • tar -zxvf xxx.tar.gz
  • -
-

check status

GPU stat:

-
    -
  • nvidia -smi
  • -
  • gpustat
  • -
-

RAM:

-
    -
  • free -h -w
  • -
-

CPU stat:

-
    -
  • top
  • -
  • htop
  • -
-

Desktop:

-
    -
  • gnome-system-monitor
  • -
- -
- - - -
- - - - - -
-
- - - - -
- - 0% -
- - - - - - - - - - - - - - - - - - - - - - - - - diff --git "a/blog/\343\200\220MATH-notes\343\200\221Differential Equation.html" "b/blog/\343\200\220MATH-notes\343\200\221Differential Equation.html" deleted file mode 100644 index d785532..0000000 --- "a/blog/\343\200\220MATH-notes\343\200\221Differential Equation.html" +++ /dev/null @@ -1,276 +0,0 @@ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -【MATH-notes】Differential Equation | 尚 Shawn's Personal Page - - - - - - - - - - - - - - -
- -
-
-
- - - - - -
- - - - - - - - -
- - - - - -
- -
- - - - - -
- -

【MATH-notes】Differential Equation -

- - - -
- - - -
-

About SDE and ODE. (building)

- - -

常微分方程(Ordinary Differential Equation)ODE

偏微分方程(Partial Ordinary Differential Equation)PDE

随机微分方程(Stochastic Differential Equation,SDE)

参考

https://youtu.be/p_di4Zn4wz4

- -
- - - -
- - - - - -
-
- - - - -
- - 0% -
- - - - - - - - - - - - - - - - - - - - - - - - - diff --git "a/blog/\343\200\220MATH-notes\343\200\221N=NP problem.html" "b/blog/\343\200\220MATH-notes\343\200\221N=NP problem.html" deleted file mode 100644 index d5d1cea..0000000 --- "a/blog/\343\200\220MATH-notes\343\200\221N=NP problem.html" +++ /dev/null @@ -1,295 +0,0 @@ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -【math-notes】n=np problem | 尚 Shawn's Personal Page - - - - - - - - - - - - - - -
- -
-
-
- - - - - -
- - - - - - - - -
- - - - - -
- -
- - - - - -
- -

【math-notes】n=np problem -

- - - -
- - - -
-

notes. (building)

- -

Abstract

It is about solving problems and finding solutions.
“How hard a problem is to solve” -Computational Complexity.

-

Turing machine(skip if you are CS student)

Keyword: Boolean->Turing->Shannon->Bell Labs->von Neumann->

-

Turing claim that there are a kind of machine which can compute any computable sequence, given enough time and memory.

-

This is the fundation of modern computer - turing machine.
Turing machine is based on boolean logic, which means datas are represented with 1&0s.
With “and”, “or”, “not”, 3 kinds of gates, you can build a system to solve problems step by step.

-

Transistors can be combined to mimic the logic gates, hence compute anything that is computable.

-

von Neumann invented the architecture of universal electronic computer, which includes input, output, CPU, memory, blabla.

-

main

Turing actually proved that not everything is computable, which left a problem for algorithms.
It turn out to be that some problem is harder than other.
Here comes the “P and NP” problem.

-
    -
  • P Problem means problems can be solved in polynomial time, is relatively easy.
    Including find shortest path, sorting a list of numbers and so on.
    These problems can be solved given enough time and compute power.

    -
  • -
  • NP Problem means problems that are easy to verify only when the solution is given.
    P problems are comtained within the NP Problems.
    For example, puzzle or Sudoku, once it is solved, it is much easier to check.

    -
  • -
-

Some previous NP problems are now studied and proved to be P problem, this lead to the question:
Are all the NP Problems really P Problems?
This determines if the computers the ultimate panacea to all problems.

-

There’s one more concept:

-
    -
  • NP Complete
    Which means all the difficult problems in NP are equivalent, and if you can prove one of those actually P Problem, all the other as well.
  • -
-

Some scientists have tried to prove that P doesn’t equal NP.
They encountered a mathematical roadblock called the Natural proofs Barrier.
The natural proofs barrier of Razborov and Rudich states that under credible cryptographic assumptions one cannot hope
to separate NP from P by finding combinatorial properties of functions that are constructive, large, and useful.
This basically tells everyone that you can’t prove P not equal NP by usual math.

-

More

Study of the natural proof barrier leads to other questions like:
How hard is it to determine the hardness of various computational problems. Why it is hard to tell how hard it is?
It opens a new field called meta-complexity.

-

参考

https://mp.weixin.qq.com/s/tjGs2XfHnQqxwsUD60J1DQ
https://cstheory.stackexchange.com/questions/30680/scope-of-natural-proofs-barrier

- -
- - - -
- - - - - -
-
- - - - -
- - 0% -
- - - - - - - - - - - - - - - - - - - - - - - - - diff --git "a/blog/\343\200\220OTHER-novel\343\200\221AI novel competition.html" "b/blog/\343\200\220OTHER-novel\343\200\221AI novel competition.html" deleted file mode 100644 index e753d77..0000000 --- "a/blog/\343\200\220OTHER-novel\343\200\221AI novel competition.html" +++ /dev/null @@ -1,288 +0,0 @@ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -【OTHER-novel】AI novel competition | 尚 Shawn's Personal Page - - - - - - - - - - - - - - -
- -
-
-
- - - - - -
- - - - - - - - -
- - - - - -
- -
- - - - - -
- -

【OTHER-novel】AI novel competition -

- - - -
- - - -
-

A little novel wrote with Claude-2-100K & DALL-E-3

- - -

不,刚刚开始

第一章

pic1
那是我逃亡的第三天夜晚。浑身酸痛,四肢乏力,我勉强驱使着忠诚的坐骑,来到了父亲的封地科林斯城堡门前。
“站住,你是谁!”夜晚值勤的士兵举起火把,警惕地喝问。我刚想开口,士兵们便重重地吸了一口凉气,惊恐地后退了几步。
“天哪!你——你是威尼斯人吗?!”
我这才注意到,逃亡在外的这几天,满身泥泞的旅行服已经被汗水浸透,头发蓬乱不堪,满脸胡茬,与我平日里的体面形象截然不同。更要命的是,我还穿着路上从一个威尼斯商人那里换来的衣服!
“等等朋友,是我,我是莫里亚总督的儿子约安!”我急忙解释道。士兵们这才认出我来,请我进了城堡大厅。
我简单地吃了些东西,开始讲述我离开君士坦丁堡以来的见闻……

-

第二章

pic2
我来到君士坦丁堡西北部的军事城镇阿德里安堡,这里的贵族对皇城的陷落还一无所知。尽管已深夜,主城堡上仍灯火通明,因为贵族们正在举办一场占星术之夜的宴会。
我见城门紧闭,只得在外面高声呼喊,请求见领主一面。过了好一会儿,一个仆从才不耐烦地过来,说今晚有重要宴会,让我明日再来。
我急切地说:“情况紧急,君士坦丁堡已被围攻多日!请你禀报领主,请求立即调兵支援。”
那仆从狐疑地看着我异域的装束,但还是进去通传了我的话。
片刻后,领主竟亲自走了出来。他穿着华丽的宴会袍服,脸上带笑,丝毫没有被我的话惊扰到。
“年轻人,今晚上罗莎小姐的占星术预见帝国国运永续长昌,你何必操这等无谓之心呢?”他打量我褴褛的衣着,饶有兴味地说。
我急切地说:“形势危急,请立即调兵支援皇城!”
领主哂笑道:“威尼斯人注定无法领会占星术的奥秘。既然星象已见好兆,我何必操劳大军前去呢?”
说着他让他的仆从给我几个钱币:“去吧年轻人,让士兵们睡个好觉,明早再启程也不迟。”
我无言以对,只得离去。就在这时,一个醉醺醺的贵族看见我的威尼斯服饰,大喝一声:“威尼斯贼人,休想夺我们的财富!”他拿起一个空盘子就向我砸来。
我躲闪不及,额头被砸出血迹。这下整个宴会乱成一团,领主不得不赶紧召来士兵维持秩序,我才得以逃离这场混乱。
科林斯的士兵们听后都大为光火,坚持要立刻率军讨伐这些昏庸的贵族。我不得不耐心地劝说他们,称我的任务是请求援军而不是讨伐叛乱。 士兵们这才不情不愿地打消了出兵的念头。然而我看得出,他们对东罗马境内的腐朽现状已经颇有微词。 这给了我很大的震撼,我自己对这一切还没有完全的看法……

-

第三章

pic3
自从皇帝与威尼斯签订协议,东罗马帝国的海军实力便日渐衰落。先前的军港塞萨洛尼基,如今已经破败不堪,码头上被废弃的军舰无人问津。
我骑马赶到塞萨洛尼基,在港口找到一艘前往目的地的威尼斯帆船。就在水手们终于提起帆准备离岸的时候,岸上突然传来一个男子的大喊:
“伙计,真为你们威尼斯人感到羡慕!这些年来,我一直在这些被抛弃的东罗马军舰上钻研造船技艺,现在我有能力建造出全世界最坚固、最宽敞、速度最快的战舰!我为自己取得的这项成就感到无比自豪!”
那人神色激动,语速极快地继续喊道:
“我永远不会效忠威尼斯或敌对势力,但是我想了很久,决定还是把我的这项发明告诉你们。这是一种划桨的新方法,为每位水手添加一个可调节的座椅,让他们可以在站立和坐下之间自由切换,这样即使减少船员人数,仍能保持更快的航速……”
“去告诉你们的威尼斯族人吧!只有你们才配拥有这项可以改变海上力量平衡的技术!”
我还没来得及解释我的身份,船已远离港口。 这项能大幅提高航速的划桨新法若落入威尼斯敌对势力之手,对东罗马帝国将是毁灭性的打击。我绝望地望着港口的景象消失在海天交际线,为未能阻止帝国重要技术泄密而深感遗憾。
我的这番叙述更加激起了科林斯士兵的义愤。“我们怎能容忍敌人窃取神圣的帝国技术!这已经是叛国的行为!” 士兵长愤怒地说。众人争先恐后地请命要去讨伐那位通敌的木工。我连忙呼吁他们保持冷静。“我们的任务是请求援军,不能擅自发动攻击。”我严肃地说。士兵们这才不情愿地按捺下怒火,然而他们的眼中已燃起复仇的火光。 我看着士兵们孤注一掷的热情,心中五味杂陈。臣民对君主忠心耿耿是美德,然而过度的狂热却会导致什么后果?我还在思索之中……

-

第四章

pic4
离开塞萨洛尼基后,我来到了雅典这座古老的城市。多年前我曾在这里求学,对这片历史灿烂的土地倍感亲切。然而现今的雅典已非我记忆中的热闹景象。街道两旁破败的石柱和荒废的市集,显露出这座城市早已失去往日的荣光。这是一条峻峭崎岖的山道,我已行走了两天一夜,口袋里的钱财早已花光。饥肠辘辘之时,我看到路边有一个农民的住所,决定去请求一些食物充饥。
我敲响了木门,一个面黄肌瘦的中年男子开了门,警惕地打量着我异乡人的装束。
“先生,求你分我一些食物充饥,我已饿了两天!” 我哀求道。
他冷笑一声:“看来我们都一样贫困!难不成你这威尼斯商人,还会比我这个佃农更穷吗?”
我还来不及解释,他已经自顾自地讲起来:“我们这些农民,为躲避皇城的苛捐杂税,只得卖掉祖传的土地,给本地贵族打工度日。都到了耳根子都要扯碎才能填饱肚子的地步,哪还有多余的食物分你们这些过路人?”
他愤怒地一脚踹向脚下的土壤:“看,这就是我现在唯一的‘口粮’了!”
我听他言辞激烈,心中也起了愤懑。然而想到自己的出身,我一个字也说不出来。只能默默转身离去,任凭饥饿继续折磨我的身体。

-

第五章

pic5
在我长时间的讲述后,士兵们都聚精会神地听完了我周详的见闻报告。我本以为他们会对当前的困境有更深刻的认识,但出乎意料的是,众人的热情丝毫未减,还是一心一意要启程去援救遥远的皇城。
我明白他们的热切来自对皇帝至高无上的忠诚,但经过旅途的所见所闻,我已经深刻意识到形势的严峻。君士坦丁堡的陷落已是定局,继续援军只会造成更大的牺牲。
于是我试图向士兵们详细解释我的判断,希望他们放弃这个盲目的计划。我建议所有人都应先休整生息,集中全力加固本地的防线,这对国家来说才是明智之举。
然而我的这番话立刻惹恼了众人。他们愤怒地挥舞着拳头,大声斥责我“怎么能在皇命有灵的关头还考虑自己的休息”。更有甚者直接拔出佩剑向我逼近,眼中满是怒火。
我惊觉自己的建议被他们视为对皇命的亵渎和对皇帝的不忠。来不及进一步解释,愤怒的士兵们已经将我牢牢制住,拖向阴森的地牢。他们用粗糙的铁链牢牢绑住我的四肢,仿佛要将我和冰冷的石壁融为一体。
“你这个来路不明的奸细,我们现在就要你好好思过!”士兵长的命令在牢房中回荡。我终于意识到,他们已经对我心存极大怀疑,我再无可能说服他们。稠密的黑暗向我压来,一切都太晚了……

-

被投入大牢后,我躺在黑暗中,脑海里不断回响士兵们对皇城危难的焦虑与盲目热情。我知道他们出于最纯粹的忠诚,但我明白那已无济于事。
我的思绪飘回多年前。 每日眺望金角湾时,我为那高高在上的圣索菲亚穹顶而自豪。直到逐渐明白,这荣光只是建立在掠夺与剥削上的虚妄。我也想起对骑士精神热血沸腾的岁月。我曾经痛恨那些被称为强盗和叛乱的穷人,并为击败他们而自豪。直到逐渐发现,他们不过是被迫反抗的农奴和失地农民,我的正义不过来自无知。最令我醒悟的,是船匠传授给我的划桨新法。他用毕生心血练就这项本可以改变海力的技艺,却因为东罗马海军的衰败而无处施展。我本可以设法保护他的心血,但最终只能眼睁睁地看着它流入威尼斯人之手。这成为我此生最深的遗憾。
种种往事在我脑海中翻滚,我终于看清这个帝国积弱的根源——暴君的腐朽、贵族的昏庸、对人才的漠视。我从盲目中解脱,知道我们需要一个崭新的秩序。
我曾经多么崇敬皇帝,认为他具有神授的威严。现在我看清他们的庸碌与腐化。神权不过是镀金的外衣,掩盖暴政的真面目。我浪费太多时间追逐虚幻的荣耀。现在,我终于理解人民的痛苦和呐喊。我们需要扫除久积的弊病,打造一个自由平等的新秩序。
我无比清晰地认识到,这正是东罗马重生的唯一道路。 即使我已无力回天,但我会带着这信念,迎接最终的审判。
在黑暗中,我坚定了重建家国的决心。这一次,我听到了历史的召唤。
就在这时,牢门打开了。光明刺入我的眼帘,士兵忧伤的脸出现在门后。
“先生,君士坦丁堡已经陷落了。”他的声音轻飘飘地传入我的耳中。
我紧紧闭上眼睛,一切都结束了。

- -
- - - -
- - - - - -
-
- - - - -
- - 0% -
- - - - - - - - - - - - - - - - - - - - - - - - - diff --git "a/blog/\343\200\220PAPER-introduction\343\200\221CausalVAE.html" "b/blog/\343\200\220PAPER-introduction\343\200\221CausalVAE.html" deleted file mode 100644 index 5ef8bca..0000000 --- "a/blog/\343\200\220PAPER-introduction\343\200\221CausalVAE.html" +++ /dev/null @@ -1,281 +0,0 @@ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -【PAPER-introduction】CausalVAE | 尚 Shawn's Personal Page - - - - - - - - - - - - - - -
- -
-
-
- - - - - -
- - - - - - - - -
- - - - - -
- -
- - - - - -
- -

【PAPER-introduction】CausalVAE -

- - - -
- - - -
-
CVPR 2021
-
- - -

CausalVAE:Structured Causal Disentanglement in Variational Autoencoder

Background knowledges

    -
  • Disentangled Representation
    seperate the representation of features into individual fractors that are independent to each other.
    The other fractors keep unchanged and the one single fractor of the generated data changes while one fractor changes.
    It aims to find a lower dimentional, explainable abstact representation of high dimentional data.
  • -
- -
- - - -
- - - - - -
-
- - - - -
- - 0% -
- - - - - - - - - - - - - - - - - - - - - - - - - diff --git "a/blog/\343\200\220PAPER-introduction\343\200\221DDPM.html" "b/blog/\343\200\220PAPER-introduction\343\200\221DDPM.html" deleted file mode 100644 index 2551081..0000000 --- "a/blog/\343\200\220PAPER-introduction\343\200\221DDPM.html" +++ /dev/null @@ -1,348 +0,0 @@ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -【PAPER-introduction】DDPM | 尚 Shawn's Personal Page - - - - - - - - - - - - - - -
- -
-
-
- - - - - -
- - - - - - - - -
- - - - - -
- -
- - - - - -
- -

【PAPER-introduction】DDPM -

- - - -
- - - -
-
Personal study notes of work DDPM.
-
- -

DDPM逐步学习

前向过程

x0->xt

-
    -
  • x0真实图像
  • -
  • xt最终噪声
  • -
  • q(xt|xt-1)以前一个状态为均值的高斯分布,从这里采xt
  • -
  • 上面一行的分布方差为beta t (t属于1-T)
  • -
  • T步数、beta 1和beta t是需要设置的
  • -
-
betas = torch.linspace(start=0.0001, end=0.02,steps=1000)
-
-
    -
  • linspace()生成随机数
    但是其实xt并不是直接通过q(xt|xt-1)得到的,还要经过重参数化
    重参数化介绍TODO
  • -
-
betas = torch.linspace(start=0.0001, end=0.02,steps=1000)
-noise = torch.randn_like(x_0)
-xt = sqrt(1-betas[t] * xt-1 + sqrt(betas[t])*noise)  # 均值和方差开根乘噪声
-
-

torch.randn_like()返回与输入大小相同的张量,填满了均值为0方差为1的随机数

-

前向过程中可以取到任意一个xt,即从这个分布q(xt|x0)采样

-
# 初始设置
-betas = torch.linspace(start=0.0001, end=0.02,steps=1000)
-alphas = 1-betas
-alphas_cum = torch.cumprod(alphas, 0) # 累乘得到alphat均值
-alphas_cum_s = torch.sqrt(alphas_cum)  # 根号alphat
-alphas_cum_sm = torch.sqrt(1-alphas_cum)  # 1-alphat
-
-# 重参数化得到xt
-noise = torch.randn_like(x_0)
-xt = alphas_cum_s[t] * x0 + alphas_cum_sm * noise)  # 均值和方差开根乘噪声
-
-

后向过程

反向的生成q(xt-1|xt,x0)是前向过程q(xt|xt-1)的后验概率分布
可以根据贝叶斯公式推导得到,并且它也是一个高斯分布,并且方差是个常量

-
# 伪代码
-betas = torch.linspace(start=0.0001, end=0.02,steps=1000)
-alphas = 1-betas
-alphas_cum = torch.cumprod(alphas, 0) # 累乘得到alphat均值
-alphas_cum_prev = torch.cat((torch.tensor([1.0]),alphas_cum[:-1]),0)
-# torch.cat将两个矩阵连接起来
-posterior_variance = betas * (1 - alphas_cum_prev) / (1 - alphas_cum)
-
-

训练目标

生成的时候我们并没有x0,那么我们要如何得到q(xt-1|xt,x0)呢?
这时候我们构造一个分布p(xt-1|xt)
它的方差和q的一致
均值则更改,将x0改为x0(xt,t),一个神经网络Unet,输入噪声图像xt和时间步t
通过一个损失函数缩小q和p之间的差距

-

最后模型无法直接通过p来生成,因为变量太多
但是,我们根据前向过程知道xt可以由x0获得,因此变换一下,就可以得到有xt得x0的式子
代入到q(xt-1|xt,x0)的均值式子我们可以得到q(xt-1|xt,x0)的均值只与xt和前向过程中t时刻加的噪声有关
所以我们修改p(xt-1|xt)的均值,将模型修改为根据xt和t预测t时刻添加的噪声,就得到了一个常量比较多的目标函数

-

训练伪代码如下

betas = torch.linspace(start=0.0001, end=0.02,steps=1000)
-alphas = 1-betas
-alphas_cum = torch.cumprod(alphas, 0) # 累乘得到alphat均值
-alphas_cum_s = torch.sqrt(alphas_cum)  # 根号alphat
-alphas_cum_sm = torch.sqrt(1-alphas_cum)  # 1-alphat
-
-def diffusion_loss(model, x0, t, noise):
-    # 根据公式计算xt
-    xt = alphas_cum_s[t] * x0 + alphas_cum_sm[t] * noise
-    # Unet预测噪声
-    predicted_noise = model(xt, t)
-    # 计算loss
-    return mse_loss(predicted_noise, noise)
-   
-for i in len(data_loader):
-    # 从数据集里读取一个batch,是真实图片
-    x0 = next(data_loader)
-    # 采样时间步
-    t = torch.randint(0,1000,(batch_size,))
-    # 生成高斯噪声
-    noise = torch.randn_like(x_0)
-    loss = diffusion_loss(model, x0, t, noise)
-    optimizer.zero_grad()
-    loss.backward()
-    optimizer.step()
-
-

采样过程

真实的推理过程必须从T步开始向前逐步生成图片
先生成一个噪声图片,每个时间步t将xt传入模型预测噪声,然后采样一个噪声
由重参数化技巧和后验概率的均值方差,可以计算xt-1,最后直到T=1

-

改进

    -
  • Improved DDPM对beta进行了改进,使得前向过程加噪声的过程更合理了一点
  • -
  • Improved DDPM还指出可以通过直接设置更小的时间步s来减小采样的时间,具体是把s与原本的t对应
  • -
-

参考

[1]微信公众号 GiantPandaCV 一文弄懂Diffusion Model

- -
- - - -
- - - - - -
-
- - - - -
- - 0% -
- - - - - - - - - - - - - - - - - - - - - - - - - diff --git "a/blog/\343\200\220PAPER-introduction\343\200\221VQ-VAE.html" "b/blog/\343\200\220PAPER-introduction\343\200\221VQ-VAE.html" deleted file mode 100644 index 78ac82e..0000000 --- "a/blog/\343\200\220PAPER-introduction\343\200\221VQ-VAE.html" +++ /dev/null @@ -1,275 +0,0 @@ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -【PAPER-introduction】VQ-VAE | 尚 Shawn's Personal Page - - - - - - - - - - - - - - -
- -
-
-
- - - - - -
- - - - - - - - -
- - - - - -
- -
- - - - - -
- -

【PAPER-introduction】VQ-VAE -

- - - -
- - - -
-

personal study note of work VQ-VAE.

- -

Background

structure

reference

苏剑林’s 科学空间(space of science)

- -
- - - -
- - - - - -
-
- - - - -
- - 0% -
- - - - - - - - - - - - - - - - - - - - - - - - - diff --git "a/blog/\343\200\220PAPER-introduction\343\200\221VideoGPT.html" "b/blog/\343\200\220PAPER-introduction\343\200\221VideoGPT.html" deleted file mode 100644 index 87e806f..0000000 --- "a/blog/\343\200\220PAPER-introduction\343\200\221VideoGPT.html" +++ /dev/null @@ -1,278 +0,0 @@ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -【PAPER-introduction】VideoGPT | 尚 Shawn's Personal Page - - - - - - - - - - - - - - -
- -
-
-
- - - - - -
- - - - - - - - -
- - - - - -
- -
- - - - - -
- -

【PAPER-introduction】VideoGPT -

- - - -
- - - -
-

personal study note of work VideoGPT()

- -

paper: VideoGPT: Video Generation using VQ-VAE and Transformers
code: github

-

Background

Multiple typies of Deep Generative Models have shown great potential and made incredible progress in last few years.
Meanwhile, high-fidelity natural videos is one notable modality that has not seen the same level of progress in generative modeling as compared to images, audio, and text.
It is reasonable because there’s high-dimentional data with spatio-temporal corral

-

structure

reference

苏剑林’s 科学空间(space of science)

- -
- - - -
- - - - - -
-
- - - - -
- - 0% -
- - - - - - - - - - - - - - - - - - - - - - - - - diff --git "a/blog/\343\200\220PAPER-summary\343\200\221Papers read 23.html" "b/blog/\343\200\220PAPER-summary\343\200\221Papers read 23.html" deleted file mode 100644 index 2bcc0fa..0000000 --- "a/blog/\343\200\220PAPER-summary\343\200\221Papers read 23.html" +++ /dev/null @@ -1,293 +0,0 @@ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -【PAPER-summary】Papers read 2023 | 尚 Shawn's Personal Page - - - - - - - - - - - - - - -
- -
-
-
- - - - - -
- - - - - - - - -
- - - - - -
- -
- - - - - -
- -

【PAPER-summary】Papers read 2023 -

- - - -
- - - -
-

papers that I’ve read in 2023.

- - - -

last edited 2023.8.11

August

8.9 Causality for Machine Learning

1
some of causality works is already in the process of entering the machine learning mainstream, in particular the view that causal modeling can lead to more invariant or robust models.
-
-Some think we are now in the middle of third revolution of information, which is also likely a conserved quantity that could only be converted or harvested rather than generating it from air.The first industrial revolutions rendered energy a universal currency; the same may be happening to information.
-One can argue that the present revolution has two components: 
-- Computer Science and AI manipulatiing symbols
-- About learning,extract information also from unstructured data
-In recent years, genuine connections between machine learning and causality have emerged, and we will argue that these connections are crucial if we want to make progress on the major open problems of AI.
-
-
2

2023.8.11 Lung nodule detection algorithm based on rank correlation causal structure learning

combine causal inference and machine learning in order to classify lung nodules.
-The whole process is defined in three parts:
-- First they obtain possible neighbor nodes(skeletons of the causal network) for both computed feature and semantic feature nodes from the input lung nodule dataset.
-- Use score based methods and algorithms to get the final causal structure network.
-- Feature selection based on the causal structure network diagram, build a subset of features by selecting computational feature nodes related to semantic fearure prediction.And train a calssifier with this new dataset to predict the level of semantic features.
-
-

2023.8.11 Toward Causal Representation Learning

From Fig.2. and Fig.3.,we can tell a relation that even a sparse change of causal variable can lead to dense like pixel level change.
-
-

July

June

May

April

March

February

January

-
- - - -
- - - - - -
-
- - - - -
- - 0% -
- - - - - - - - - - - - - - - - - - - - - - - - - diff --git "a/blog/\343\200\220PHIL-guide\343\200\221What Causality tells in real life.html" "b/blog/\343\200\220PHIL-guide\343\200\221What Causality tells in real life.html" deleted file mode 100644 index 5482688..0000000 --- "a/blog/\343\200\220PHIL-guide\343\200\221What Causality tells in real life.html" +++ /dev/null @@ -1,410 +0,0 @@ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -【PHIL-guide】What Causality tells in real life | 尚 Shawn's Personal Page - - - - - - - - - - - - - - -
- -
-
-
- - - - - -
- - - - - - - - -
- - - - - -
- -
- - - - - -
- -

【PHIL-guide】What Causality tells in real life -

- - - -
- - - -
-
What Causality tells in real life.
-
- -

注:以下内容很多废话,没有严谨的推理,数据、理论来源很杂,历史的长河中肯定已经诞生过同构的想法,这一篇主要是拾人牙慧、繁辞多言而已。

-

第 零 章===

晋平公十九年(前539年),晏子与叔向在宴会上交谈了起来,晏子感叹道:“齐国大概要归于陈氏了。” 叔向听了这话也感慨万千,虽然晋国此时仍在晋公室治理之下,
虽然自己也或多或少得益于这个体制,但是国内百姓疲敝不堪,公室成员敛财不止,道路上饿死了吗的人相遥望,绛城独生女屋里的财物无处摆放,
百姓听到新的政令颁布就像躲避仇人一样…自己的羊舌氏势力孱弱,也没有好的子嗣,自己该怎么办?

-

第 一 章===

(摘自正在编写的另一篇文章-汉代剑客如何提升编程能力?,全文将会在后续单独发出)

-

在增长工程师Phodal Huang的个人博客里,大佬提到了一种工作时的技术提高方式

-

Fig.1

-

在这里暂不讨论武术(及其他东西)的归属性。

-

我们发现,原来学编程,还能用学习武术的方式?

-

与此同时,英国一瘦弱老头Matt发现了一件古怪的拆尼斯兵器:

-

Fig.2

-

据称这件奇怪(奇怪吗)的兵器来自于一个私斗盛行的时期,属于遥远的汉朝。稍有常识的人都能发现,这种有着神似欧洲的Saber军刀护手的兵器,包括其可能代表的决斗风气一同早了欧洲(为人熟知的形象(看出求生欲了吗?))千年。

-

不讨论刀剑形制的归属性。我们发现,原来一直以无护手、小护手形象出现的中国刀剑,居然有独立研发出的实用哆啦A梦护手,并且路子和千里之外欧洲人的saber基本一样?

-

是什么将武术与计算机连接了起来,将汉朝剑客与西方剑士们连接了起来?

-

All that you touch,

All that you see,

All that you tase,

All you feel

以眼睛对外部世界的感知为例,视网膜上的光感受器(视杆和视锥)在感受到光之后产生局部分级信号,随后光感受器影响双极细胞产生分级信号,
-
-

神经节细胞又受到双极细胞产生分级电位。最终神经节细胞的分级电位达到一定程度后使得细胞膜去极化超过一定水平产生动作电位,
视觉刺激越强烈,动作电位的频率越高。动作电位先是传到丘脑的外膝核,随后是更高级中枢,最终产生感知。外膝核的轴突通过视放线投射到大脑皮层,
皮层内部来自不同外膝核神经元的信息会进一步汇聚和分散。。。。。。

-

剑客们本身是这个世界里一个个活生生的生物,他们都拥有亿万斯年进化赋予的双眼、耳朵、鼻子,这些让他们可以轻易地获取许多信息。但是显然这并不足以创造出哆啦A梦护手。

-

……

-

稍作整理,我们发现,这些东西的中间件是人,尤其是人脑。

-

这些现象都是人类在面对问题与困难(生或死、人口红利或社会人才)时,发挥聪明才智实现的。可能正因为我们拥有类似的思考方式,类似的“智能”,所以针对一个类似的问题我们可以反复独立造出同样的轮子。

-

可以想象,汉代的剑客们看到了无数“下殿数交, 三中其臂, 左右大笑”的断臂维纳斯故事后,对持剑决斗这一场景进行了“抽象”,“建模”,“量化”,“求解”给出了手与臂易受伤这一问题的解决方案,同时不停地在实战中根据反馈修正,最后得到了这个最好用的铁盘一样的护手,西方的剑士大概也经历了同样的过程,得到了类似的成果,只不过,哆啦A梦护手在我们这里并没有产生大的影响,原因已不可知。。。

-

第 二 章===

荒原

大楼黑色的身影边缘,云层呈现深红色。北京又开始下雨了。

-

近期,甚至说近年来,肯定也有很多人和我一样能够感受到社会的戾气水平(以暴力作为解决问题的第一手段的概率)正在以一个难以置信的速度升高,每个人心里都有一团怨气。

-

Fig.3

-

秦人,其生民也狭厄,其使民也酷烈。

-

社交平台上,关于一个无关紧要的问题的一点点的分歧可以引来无数人追着骂几百层,几乎所有评论的目的都不是说服对方,而是想尽一切办法让对面破防;
种种在事外者看来无所谓的事件,如商家不小心给顾客戴上了手镯拿不下来,都可以造成当事人极大的情绪应激;出现大量的帖子劝诫大家谨慎花钱,生活中遇到小摩擦主动退让,而另一批人劝大家遇到小摩擦一定要摆出闹事者的态度;对“精神状态”的担心成为出现频率极高的问候;躺卷韭润四学的热度越来越高。

-

Fig.4

-

我们正在经历什么?

-

回顾一下我这20来年来的人生中,发生了些什么。

-
    -
  • “1998年至2000年, 中国国有企业共发生下岗职工2137万人。” 迎来人口红利。

    -
  • -
  • 2001年,中国加入了世界贸易组织,迎来全球化的广阔市场之红利。

    -
  • -
  • 中国GDP常年以10%以上的速度快速增长,在极短的时间里超越了日本成为世界第二大经济体。

    -
  • -
  • 2008年11月,针对世界性的金融风暴,我国推出了一系列措施应对出口受影响导致的经济增速下滑,4万亿流入了各个行业,尤其是基建。

    -
  • -
  • 2012年,我国15-59岁的劳动年龄人口首次下降。

    -
  • -
  • 2013、2014年,房地产不景气,主要城市房价出现集体下跌,资本急于出走,对经济和就业产生冲击。十年国债收益率已经升至历史高位水平。银监会批准了很多民营金融机构的建立。

    -
  • -
  • 一带一路提出。

    -
  • -
  • 2015年,我国开始推行金融创新,房地产呈现金融化趋势。此前A股态势一直呈现大牛市,直到2015年6月中旬,一轮暴跌扑面而来,被称作2015中国式股灾。

    -
  • -
  • 随着4G推广完成以及手机行业的发展,互联网行业蒸蒸日上。B站在2015年产出了大量鬼畜视频,如今很多还以数千万播放量存在于播放量排行榜首页,其中出圈的有”Duang”等。

    -
  • -
  • 2016年,中国居民债务收入比突破100%。

    -
  • -
  • 二胎放开。

    -
  • -
  • 字节跳动孵化了一款软件-抖音,随后短视频开始出现在人们视野。

    -
  • -
  • 至2016年下半年,陆续有两万亿美元左右的资金流出至境外。

    -
  • -
  • 2017年,B站鬼畜区出现了大量怀念2015年鬼畜区的视频。

    -
  • -
  • 决定设立河北雄安新区。

    -
  • -
  • 2018年,居民债务收入比从2015年的89%上涨到了121%。

    -
  • -
  • p2p大量暴雷

    -
  • -
  • 房住不炒

    -
  • -
  • 之后的事情大家都熟悉,贸易战,疫情

    -
  • -
  • 2019年3月,996.ICU事件爆发,网上对劳动法、加班的讨论越来越多。

    -
  • -
  • 居民债务收入比继续从128%升至140%以上

    -
  • -
  • 中国总和生育率迅速下降

    -
  • -
  • 2020年,贵州独山县走入大众视野

    -
  • -
  • 2020年9月,一篇名为《外卖骑手,困在系统里》的文章刷屏

    -
  • -
  • 恒大暴雷

    -
  • -
  • 2022年,中国劳动人口/老年人口比已经跌至日本90年代初水平

    -
  • -
  • 地方债务率已达到120%

    -
  • -
  • 2023年8月,中融信托暴雷

    -
  • -
  • ……

    -

    1820年至1870年,美国城市化率不断升高由7%上升到25%;1847年,梭罗推出了他的《瓦尔登湖》;

    -

    对于我们父母那一辈人来说,他们的青春年华1978至2000年的二十年非常波澜壮阔,这二十年间中国城市化率上升了约20%。很多人抱着和马督工一样的想法,踏上寻找自己的天空之城的道路。

    -

    1860s,南北战争爆发;此后第二次工业革命曙光初现,美国城市化率开始快速提高。1876年,《汤姆索亚历险记》发表。至1930年代,美国城市化率提高了约30%,达到了56%。1936年,《飘》出版;

    -

    新世纪的二十年同样波澜壮阔,中国城市化率继续上升了约30%,无数的小镇青年选择或被选择,独自或举家,前往北上广深。有的人留下来了,有的人还在挣扎,有的人只能面带笑容灰溜溜地回家。

    -

    1830s至第一次世界大战前,德国通过两次工业革命极大地推动了工业化,城市化率从20%以下升至约80%。马克思等人对此时的德国进行了细致的分析,包括指出软弱的德国工商业小市民在改良的道路上缓慢地试图向大资产阶级转化。

    -

    这数十年间我们所处的环境剧烈变化了,一线大城市虹吸了大量的年轻人,直到把县城和农村变为荒原。城投债,地方债,GDP货币化,双轨制养老金,影子银行……这数十年间,每个时刻每个城市的每个角落中都在发生着这样的故事:

    -

    国企银行用低息将居民变为债务人,用这些存款投资催动基建计划,推着无数的农民工建起了一座座不属于他们在内的任何人的大厦,无数的血汗工厂工人、骑手、快递员、服务员、司机用体力、专注力服务一个个顾客, 创造出服务价廉的“方便”天堂,为了能够多多加班赚一些微薄的银两,他们已经无法、无暇用亿万斯年的进化赋予他们的美丽双眼去观察身边的世界,只能在晚上通过互联网看看其他阶级享受被构建出的生活方式暗下决心后匆匆睡去,最大的理想是“不生病”;

    -
  • -
-

Fig.5

-

无数的人怀着对美好生活的向往挣扎着成为城市居民。一些人相信可以通过老老实实一步一个脚印赚到钱,寻觅许久找到便宜的地方开了商铺成为个体户,面对棚改被赶到了商业中心交着数倍的地租,不得不让位于预制菜或者开始和城管游击战; 一些人挣扎求生刷题考证堆笑脸堆热情,成为了小小公务员、企业职员、教师、自由职业者、普通医生律师工程师,抱着自己的工作还有一些社会价值,或者还有可能成为全球化到处飞的精英白领最后实现财务自由的想法,塌着背坐在办公桌前写着本子,做着与自己剩余部分的人生毫无关系的工作,想到十年白天最美好的8+个小时的时间就这样流走,他们不得不或把自己沉浸在琳琅满目的商品中或多或少地变成恋物癖,成为上面那些人的忠实客户,用皮包和手表标识一下自己的生活方式,或在下班重新梳洗打扮,在dating app或夜店寻求同事外的社会关系,或花几十万在家里装修营造一个小窝,放置满喜欢的家具手办直到一场大火发生,或在小说、综艺、动漫、荒郊野岭寻求一个美好的异邦放置自己的闲暇,或深陷针对人的生理奖励机制特别设计的商品化爱好冲浪滑雪速降LiveHouse,在周末结束后感叹自己永远也不要生孩子,另外其余部分的生活怎么会如此不像是活着。。。

-

Fig.6

-

闷头卷上来的小企业主、处级县处级干部、教授、中级经理、小有名气的工程师医生、二三线明星,可能已经数年没有升迁或涨工资,可能已经不再幻想自己终有一天可以追得上上车早的那些人,取得同样的成就,变成大企业家银行家、走进省部甚至大会堂、成为富商,成为操控一个区域一个城市一个行业的人。自己已经到了无法停下脚步改换赛道的年龄,不论是柜子后备箱放满了烟酒还是平时只在食堂吃点馒头,这些人往往面临着社会中最棘手的一些问题,也许会逐渐在承受四面八方压力的过程中,慢慢麻木,只能靠自己经历训练形成的思维模式做一些看起来没有任何问题的决策,用着已经变形的身体挤出的最后的一丝力气,去从精神上关注家里的孩子。

-

大家都清楚自己经历了什么,但是国民党的士兵投诚后还需要解放军来组织诉苦大会,才能最终明白原来自己所经历的一切不是因为什么命不好,前世没有积德。

-

所有人生活中都越喊越响,但是在某种奇妙的氛围中大家全部都政治性养胃了。

-

……

-

第 三 章===

    -
  • “我已对哲学、法学、以及医学方面,还对神学都花过苦功,彻底钻研,我被称为博士,牵着学生们的鼻子,上上下下,纵横驰骋”……“枯守这个牢笼,在这里暗淡无光”……”唉!真想披着你的银光, 漫步在高高的山顶上, 随精灵们绕着山洞飞旋;在洒满你幽辉的草地徜徉, 扫净知识的雾障,沐浴你的清露, 还我身心健康!”
  • -
-

-

一旦我甘心躺上软床, 那就随便你把我怎样!你要能甜言蜜语哄我, 使我自满和孤芳自赏, 你要能使我沉溺享乐, 那我的末日就已到啦!这就是我打赌的筹码!

-

-

我愿看见这样熙熙攘攘的一群—— 在自由的土地立足的自由之民。那时对眼前的一瞬我便可以说:你真美啊,请停一停!

-

《浮士德》

-

*

-
我从小就很喜欢历史,但也一直在想,这一柜子书好像大家看来全是闲书,历史到底有什么意义?
-
-“知兴替”如果只解读成了解兴替的故事,那历史真的是太没用了。就像一个相机、一个个摄像头、一双眼,如果只是看到信息,那真是太没用了。
-
-有一天你火急火燎走上地铁,想找一个位置坐下,可惜没有空位了,于是我们进入了一个解决问题的过程。我们开始观察,第一个座位坐的是个男生,在刷手机,第二个位置是个老人,抱着行李箱......你观察了这个车厢所有的人,甚至都猜测了他们今天经历了什么、心情如何。最后你锁定了一个最有可能下车的人,因为他看了看下一站位置,开始收拾东西。于是你站了过去。过了一会,他确实起身了,但是没想到,在你坐下之前,旁边座位的男生挪了过来占据了这个位置,因为他和女朋友之前被起身的人隔开了。当然,站你旁边的人就这样坐到了男生原先的位置。
-
-在这个过程中你收集了很多信息,没有做错任何事,但最后还是没有解决问题。
-
-

我越来越相信,数据、信息确实很重要,这几十年来统计学催生出很多重要且有趣的事物,但是因果还是比数据要有力得多。不局限于双眼的观测,我们还可以通过对现实的直接干预,比如直接开口询问其他乘客何时下车,确认能否让座,来进一步获取信息进行决策;坐下之后,我们可能还会继续想,如果不是我刚才脸皮厚了点,可能就要继续站着了。最后,我们可能会弄清楚,原来成功坐到座位这件事背后需要考虑的因果逻辑因素有坐着的人是不是准备下车,还有旁边的人会不会突然挪位置等等很多。

-

这不是一个太恰当的例子,但是我想说明的是,一个可以说千百年来无数人说烂了的点,请务实乎。

-

回到第一章,想象一下打剑时遇到什么样的对手最轻松?不是单纯柔弱体能差的,因为刀剑本身就磨平了人与人之间的力量差距;不是门外汉无章法的,这种人捉摸不透反而容易翻车;我想,最理想的对手就是那种典型的稚嫩新手,老乌龟一样慢腾腾的,拿着武器不动弹,甚至有点不敢打,死死盯着你,只会等着你打过去的时候防御的人,这种人让我联想到社恐。稍微晃一下这些人就被骗,进攻完也不用担心被对方反击,只要看准下一个空继续进攻就好。一对一的博弈对这些只有双眼时刻观察的人来说尚且如此,多人在线PVP的社会化大熔炉对这些人来说,简直算得上地狱了。

-

回到第二章,我不是完全的讨厌,但是我不是很喜欢这样的社会,并且我愈发的觉得,我们应该真的坐下来,不是说地铁座位,是说找到一个闲暇的时刻,静静地坐一会,把遇到的问题真的当做一个问题来看。抛开那些“ChatGPT可以解决一切blabla”“解放发展生产力可以解决一切blabla”“力大砖飞”“干就完了”“看着这个预言,我们要准备好这样做”……,我们从观察美丽的世界开始,找到对应问题对应环境的因果变量、发掘一些因果关系;然后是干预改变世界,包括这么做会怎么样,怎么办;最后是反事实,假如我们没有这么做,会怎么样?这样我们才能找出无论是正确的还是错误的真正原因,乃至宣告问题得到解决。人类受益于这样的能力已经很久了,只不过我们面对的问题也越来越复杂越来越难以解决。

-

现在我坚信行动带来的东西一定更多,干预会让我们发现更多更全面的因果链条,让我们掌控而不是被动地面对现实,就像迅猛地侵略直击对手的面部,紧紧地贴住敌人,对手反而无法挥动他的长刀;我坚信就算是慢腾腾的新手,也能成长为厉害的剑士,只要他挨打的时候知道疼,知道丢人,了解基本技法之后懂得出手进攻并学习更多的情况,只要他内心还是想要掌握这项技能,想成为一个剑客,那么“那隐蔽着的宇宙的本质自身并没有力量足以抗拒求知的勇气。对于勇毅的求知者,它只能揭开它的秘密,将他的财富和奥妙公开给他,让他享受”。

-

现在终于让我们回到“历史的意义”以及《怎么办?》(Что делать?)。

The School of Life ,一个YouTube频道认为,研究历史的目的是为了发掘某些被掩盖在历史长河中的,好的,有启发性的idea,并且将它们用来解决我们自己的时代的问题。

-
   1469年,20岁的Lorenzo de' Medici接管了美第奇家族的银行,他延续美第奇家族之前对画家、雕塑家等艺术家、哲学家的赞助,他甚至邀请了很多艺术家住进他的家里。同时他扩建了家族的美第奇图书馆,大量派人搜寻古代的书籍,寻找有趣的理念。这一切都是以促进美、真理和智慧的慈善行为推动的。
-
-  绘画、雕塑,再加上考古、哲学等等各类科目,也和历史一样是“意义缺失”的重要怀疑对象。回头看我们会发现文艺复兴恰恰是围绕着这些进行的,人们发掘出古希腊人某种积极对待生活的哲学、绘画表现生命之短暂、设计出完美平和的庭院,让妈妈的声音可以从窗户上传递到广场另一边的孩子处,宣扬统治者德行的重要性...最最重要的就是,他们首先想要*成功地管理他们的社会*,为此他们想要让人民开化、建造宏伟的城市、让所有人因平和、美丽、悦人的公共生活而感到振奋。欧洲的银行家和金融家们用他们的行动部分促成了这一场带有目的性的文艺宣传——对他们所处社会环境的一次直接干预。这些艺术品背后蕴含的对美、公正社会制度的追求,让苏联时期遗留的雕像与画作与独山县的大楼划清了界限,让圣母玛利亚和时装品牌模特得以区分开来,让Pink Floyd的Time与单纯让人听了睡不着觉的歌摆放在不同的分区。
-
-

Fig.7

-

孩子一眨眼就长大,随后老去死亡,”The time is gone the song is over,thought I’d something more to say…”

-

我们没有多余的时间花费在争吵和攻击别人上,我们珍贵的专注力更是应该用在解决困难的问题上。
Marsilio Ficino提出一种想法,即我们的美丽、性感,对金钱、名誉的欲望,应该用在对崇高的智慧的追求。显然不是像今天一样用在对印了大字母的手提袋子和玻璃瓶装的香水的追求上。
(因为没有什么中文信息,我翻译了Marsilio Ficino的Wiki页面,并将在后续继续(随缘)更新一些国内找不太到资料的人物:https://captaincccp.github.io/2023/08/04/Marsilio%20Ficino/
如果一个人一个月5000元,那么6万元钱可以养活一个人1年,54万元可以养活3个人3年。我不是在为任何人给出做慈善的建议,阻拦艺术创作的因素也不止有钱,我只是想说,相比于北上的房价甚至很多人的流动资产来说,对我们自己所处的世界做一些干预的代价并没有想象的那么多。我们应该对我们应当如何花费自己的钱财,如何花费自己剩余的时光,如何使用自己学习到的知识进行创造,有更多的认识,并且我真切地希望我们可以通过各个学科合作的,有目的性的,专注地,全心全意实事求是地,探求社会运转的本源逻辑,来改善我们的生活质量,创建宜居的城市,实现“自由的人民在自由的土地立足”。
Fig.8

-
                                                                   初写于2023年7月10日
-
- -
- - - -
- - - - - -
-
- - - - -
- - 0% -
- - - - - - - - - - - - - - - - - - - - - - - - - diff --git "a/blog/\343\200\220TRANS-work\343\200\221Marsilio Ficino.html" "b/blog/\343\200\220TRANS-work\343\200\221Marsilio Ficino.html" deleted file mode 100644 index 39db461..0000000 --- "a/blog/\343\200\220TRANS-work\343\200\221Marsilio Ficino.html" +++ /dev/null @@ -1,430 +0,0 @@ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -【TRANS-work】Marsilio Ficino | 尚 Shawn's Personal Page - - - - - - - - - - - - - - -
- -
-
-
- - - - - -
- - - - - - - - -
- - - - - -
- -
- - - - - -
- -

【TRANS-work】Marsilio Ficino -

- - - -
- - - -
-

Translation for English version wiki, 2023.8.4
以下内容翻译自维基百科Marsilio_Ficino,除翻译外未经修改。
根据CC BY-SA 4.0协议,可以自由复制
The Renaissance –Marsilio Ficino

- -

Marsilio Ficino

    -
  • Marsilio Ficino (意大利语: [marˈsiːljo fiˈtʃiːno]; 拉丁名: Marsilius Ficinus; 1433年10月19日 – 1499年10月1日)。意大利学者和天主教牧师,意大利文艺复兴早期最有影响力的人文主义哲学家之一。
  • -
  • 他是一位占星家,是新柏拉图主义的复兴者,并与当时的学术界保持联系,也是第一位将柏拉图现存完整著作翻译成拉丁语的译者。[2]
  • -
  • 他的佛罗伦萨学院是复兴柏拉图学院的尝试,影响了意大利文艺复兴的方向和基调以及欧洲哲学的发展。
  • -
-


Domenico Ghirlandaio在 佛罗伦萨 新圣母玛利亚托尔纳博尼教堂 绘制的壁画中的Marsilio Ficino

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
内容
出生1433年10月19日 佛罗伦萨共和国 费格林法尔达诺(Figline Valdarno)
逝世1499年10月1日 (65岁) 佛罗伦萨共和国 卡雷吉(Careggi)
著名工作1474 De Christiana Religione
1482 Theologia Platonica, de immortalitate animae
1484 De amore
1489 De vita libri tres
亲属Diotifeci d’Agnolo, Alessandra di Nanoccio (父母)
时代Renaissance philosophy
学派Christian humanism
Neohermeticism
Neoplatonism
Augustinianism
Thomism
主要领域Immortality of the Soul
Theology of Love and Eros
Translation of Platonists
Commentary on the dialogues of Plato
Natural, medical use of Astrology and Theurgy
Catholic interfaith polemics
著名思想Platonic love
First in a genus
Prisca theologia[^1]
受影响于单元格
影响于单元格
-

早期生活

    -
  • 菲奇诺出生于菲利内瓦尔达诺。 他的父亲迪奥蒂菲西·达尼奥洛 (Diotifeci d’Agnolo)
    是科西莫·德·美第奇 (Cosimo de’ Medici) 赞助下的一名医生。
  • -
  • 科西莫·德·美第奇 (Cosimo de’ Medici)
    将这位年轻人带入自己的家庭, 成为马西里奥 (Marsilio) 的终身赞助人,
    而马西里奥则成为他的孙子洛伦佐·德·美第奇 (Lorenzo de’ Medici) 的家庭教师。
  • -
  • 意大利人文主义哲学家和学者乔瓦尼·皮科·德拉·米兰多拉(Lorenzo de’ Medici. Giovanni Pico della Mirandola)是他的另一位学生。
  • -
-

事业与思想

柏拉图学院

    -
  • 1438年至1445年,佛罗伦萨的费拉拉-佛罗伦萨会议中,
    在弥合东方(东正教)和西方(天主教)教会之分裂的失败尝试期间,
    科西莫·德·美第奇和他知识界的朋友们结识了新柏拉图哲学家乔治·杰米斯托斯·普莱松 (George Gemistos Plethon),
    他对柏拉图和亚历山大神秘主义者思想的论述令佛罗伦萨的人文主义者着迷,以至于他们称他为第二个柏拉图。

    -
  • -
  • 1459 年,约翰·阿吉罗普洛斯 (John Argyropoulos) 在佛罗伦萨讲授希腊语言和文学,菲西诺成为他的学生。[3]

    -
  • -
  • 当科西莫决定在佛罗伦萨重建柏拉图学院时,他选择菲西诺作为院长。
    1462年,科西莫向菲奇诺提供了柏拉图著作的希腊语手稿,
    随后菲奇诺开始将整个文集翻译成拉丁文[4](对话草稿的翻译于1468-1469年完成;[5]于1484年出版)。

    -
  • -
  • 菲奇诺还翻译了列奥纳多·达·皮斯托亚(Leonardo da Pistoia)发现的希腊化时期的希腊语文献集(后来称为赫尔墨斯)[6],
    以及许多新柏拉图主义者的著作,包括波菲利(Porphyry)、扬布利库斯(Iamblichus)和普罗提诺(Plotinus)。

    -
  • -
  • 他的众多学生包括尼科洛·瓦洛里[7][8]和弗朗西斯科·卡塔尼·达·迪亚塞托。
    后者被菲西诺认为是他佛罗伦萨柏拉图学院院长位置的继任者。[9]
    迪亚塞托的学生乔瓦尼·迪·巴尔多·科西 (Giovanni di Bardo Corsi) 于1506年创作了一部菲奇诺的传记。[10]

    -
  • -
-

神学、占星学和灵魂

    -
  • 虽然接受过医生培训,但于菲奇诺还是在1473年成为一名牧师。[11][12][13]
    1474年,菲奇诺完成了关于灵魂不朽的论文《Theologia Platonica de immortalitate animae》[3] (Platonic Theology)(《柏拉图神学》)
    和《De Christiana Religione》(On the Christian Religion)(《论基督教》),一部宗教史和对基督教的维护。[14]
    每一次对古典时期激情地再次探索,都令他对占星学表现出了更多兴趣(尽管他诋毁其与神启的关系),
    这使他陷入了与天主教会的麻烦。1489年,他在教皇英诺森八世面前被指控为异端[3],随后被无罪释放。
  • -
-

菲奇诺在1492年写道:

-
-

“这个世纪,简直是黄金时代,几乎绝迹的文献重现光明:文法、诗歌、修辞、绘画、雕塑、建筑、音乐……这个世纪似乎让占星术更完善了。”

-
-

菲奇诺1474年至1494年期间的信件至今仍然得到保存并且已得到出版。[3]
他于1484年撰写了《De amore》(Of Love)(《爱情》)。
《De vita libri tres》(Three books on life)(《关于生命的三本书》)
或《De triplici vita》[15] (The Book of Life)(《生命之书》)于1489年出版,
提供了大量关于保持健康和活力的医学和占星学建议,并拥护新柏拉图主义关于世界灵魂(world’s ensoulment)及其与人类灵魂融合的观点:

-
-

有一些迷信而盲目的人,他们能够在即使是最低等的动物和最卑微的植物身上看到生命,却无法感受到天堂或整个世界的生命…
如果现在那些人还承认世界上最微小的粒子也具有生命,那是多么的愚蠢!多么的嫉妒!
他们既不了解我们所生活、运动和存在的世界本身就是有生命的,也不希望事实是这样。

-
-

这种统一的“生命力”思想的一个代表就是菲奇诺的占星学。在《生命之书》中,他详细描述了行为与后果之间的内在联系,
给出了一系列支配一个人命运的事情。

-

医学工作

    -
  • 可能是由于他的父亲迪奥蒂费奇(科西莫·德·美第奇的医生)在早期的影响,
    菲奇诺发表过医学主题的拉丁语和意大利语论文,例如
    《Consiglio contro la pestilenza》(Recommendations for the treatment of the plague)(鼠疫治疗建议)
    和 《De vita libri tres》(Three books on life)(关于人生的三本书)。
  • -
  • 他的医学著作对帕拉塞尔苏斯(Paracelsus)等文艺复兴时期的医生产生了相当大的影响,
    他曾与帕拉塞尔苏斯等人分享了微观宇宙和宏观宇宙的统一,及其间生理和心理形式的相互作用的看法,并试图研究它们的特征来治疗疾病。
    这些涉及占星学和炼金术概念的作品在当时非常流行,以至于菲奇诺被怀疑为异端,特别是在1489年内含 如何在这个包含恶魔和其他灵魂的世界中健康生活 的具体说明的第三本书出版之后。[17]
  • -
-

柏拉图式恋爱

    -
  • 值得注意的是,菲奇诺创造了“柏拉图式爱情”这个概念,该词首次出现在他1476年写给阿拉曼诺·多纳蒂(Alamanno Donati )的信中。
    1492年,菲奇诺出版了《Epistulae》(Epistles)(《书信》),其中包含写给他的学术同事和终生挚友乔瓦尼·卡瓦尔康蒂(Giovanni Cavalcanti)
    的用拉丁语写成的柏拉图式情书,其中谈到了柏拉图式爱情的本质。 因此,有人声称菲西诺是同性恋,但这在他的信件中找不到依据。[18]
    不管怎样,菲奇诺写给卡瓦尔康蒂的信使“柏拉图式爱情”一词在西欧流行起来。
  • -
-

死亡

    -
  • 菲西诺于 1499年10月1日在卡雷吉去世。
    1521 年,为了纪念他,安德烈亚·费鲁奇 (Andrea Ferrucci) 为他雕刻了一座半身像,
    这座半身像位于圣母百花大教堂中殿的南侧
  • -
-

出版

    -
  • Theologia Platonica de immortalitate animae (Platonic Theology). Harvard University Press, Latin with English translation.
      -
    • vol. I, 2001. ISBN 0-674-00345-4
    • -
    • vol. II, 2002. ISBN 0-674-00764-6
    • -
    • vol. III, 2003. ISBN 0-674-01065-5
    • -
    • vol. IV, 2004. ISBN 0-674-01482-0
    • -
    • vol. V, 2005. ISBN 0-674-01719-6
    • -
    • vol. VI with index, 2006. ISBN 0-674-01986-5
    • -
    -
  • -
  • The Letters of Marsilio Ficino. Shepheard-Walwyn Publishers. English translation in 11 vols. with extensive notes; the Language Department of the School of Economic Science.
  • -
    • -
    • vol. I, 1975. ISBN 978-0-85683-010-5
        -
      • vol. II, 1978. ISBN 978-0-85683-036-5
      • -
      • vol. III, 1981. ISBN 978-0-85683-045-7
      • -
      • vol. IV, 1988. ISBN 978-0-85683-070-9
      • -
      • vol. V, 1994. ISBN 978-0-85683-129-4
      • -
      • vol. VI, 1999. ISBN 978-0-85683-167-6
      • -
      • vol. VII, 2003 ISBN 978-0-85683-192-8
      • -
      • vol. VIII, 2010 ISBN 978-0-85683-242-0
      • -
      • vol. IX, 2013 ISBN 978-0-85683-289-5
      • -
      -
    • -
    -
  • -
  • Commentaries on Plato. I Tatti Renaissance Library. Bilingual, annotated English/Latin editions of Ficino’s commentaries on the works of Plato.
      -
    • vol. I, 2008, Phaedrus, and Ion, tr. by Michael J. B. Allen, ISBN 0-674-03119-9
    • -
    • vol. II, 2012, Parmenides, part I, tr. by Maude Vanhaelen, ISBN 0-674-06471-2
    • -
    • vol. III, 2012, Parmenides, part II, tr. by Maude Vanhaelen, ISBN 0-674-06472-0
    • -
    -
  • -
  • Icastes. Marsilio Ficino’s Interpretation of Plato’s Sophist, edited and translated by Michael J. B. Allen, Berkeley: University of California Press, 1989.
  • -
  • The Book of Life, translated with an introduction by Charles Boer, Dallas: Spring Publications, 1980. ISBN 0-88214-212-7
  • -
  • De vita libri tres (Three Books on Life, 1489) translated by Carol V. Kaske and John R. Clarke, Tempe, Arizona: The Renaissance Society of America, 2002. With notes, commentaries, and Latin text on facing pages. ISBN 0-86698-041-5
      -
    • “De triplici vita”. World Digital Library (in Latin). 16 September 1489. Retrieved 1 March 2014.
    • -
    -
  • -
  • De religione Christiana et fidei pietate (1475–6), dedicated to Lorenzo de’ Medici.
  • -
  • In Epistolas Pauli commentaria, Marsilii Ficini Epistolae (Venice, 1491; Florence, 1497).
  • -
  • Meditations on the Soul: Selected letters of Marsilio Ficino, tr. by the Language Department of the School of Economic Science, London. Rochester, Vermont: Inner Traditions International, 1996. ISBN 0-89281-658-9. Note for instance, letter 31: A man is not rightly formed who does not delight in harmony, pp. 5–60; letter 9: One can have patience without religion, pp. 16–18; Medicine heals the body, music the spirit, theology the soul, pp. 63–64; letter 77: The good will rule over the stars, p. 166.
  • -
  • Commentary on Plato’s Symposium on Love, translated with an introduction and notes by Sears Jayne. Woodstock, Conn.: Spring Publications (1985), 2nd edition, 2000. ISBN 0-88214-601-7
  • -
  • Collected works: Opera (Florence,1491, Venice, 1516, Basel, 1561).
  • -
-

相关

    -
  • Allegorical interpretations of Plato
  • -
  • Contemporary Italian Renaissance philosophers:
  • -
  • Lodovico Lazzarelli
  • -
  • Giovanni Mercurio da Correggio
  • -
  • Giovanni Pico della Mirandola
  • -
  • Greek love
  • -
  • Hermetica (philosophical writings attributed to Hermes Trismegistus)
  • -
  • Hermeticism
  • -
  • Platonic Academy (Florence)
  • -
  • Platonic love
  • -
  • Renaissance humanism
  • -
  • Renaissance philosophy
  • -
  • Renaissance magic
  • -
  • Translations (historical)
  • -
-

引用

[1]:Heiser, James D., Prisci Theologi and the Hermetic Reformation in the Fifteenth Century, Repristination Press, 2011. ISBN 978-1-4610-9382-4
[2]:Marsilio Ficino. Voss, Angela. Berkeley, Calif.: North Atlantic Books. 2006. pp. ix–x. ISBN 1556435606. OCLC 65407018.
[3]:One or more of the preceding sentences incorporates text from a publication now in the public domain: Symonds, John Addington (1911). “Ficino, Marsilio”. In Chisholm, Hugh (ed.). Encyclopædia Britannica. Vol. 10 (11th ed.). Cambridge University Press. pp. 317–319.
[4]:Bartlett, K. R., ed. (2011). The Civilization of the Italian Renaissance: A Sourcebook. University of Toronto Press. ISBN 978-1442604858.
[5]:Hankins, J. (1990). Plato in the Italian Renaissance. p. 300. ISBN 9004091610.
[6]:Yates, Frances A. (1964) Giordano Bruno and the Hermetic Tradition. University of Chicago Press 1991 edition: ISBN 0-226-95007-7
[7]:Nuovo Dizionario Istorico, Va = Uz, Tomo XXI, translated from French, Remondini of Venice (1796); pages 51.
[8]:Niccolo Valori (died 1527) wrote a biography of Lorenzo de’ Medici the elder and published posthumously in 1568.
[9]:Marsilio Ficino, entry by Christopher Celenza in the Stanford Encyclopedia of Philosophy
[10]:Annotated English translation of Corsi’s biography of Ficino Archived 15 October 2011 at the Wayback Machine
[11]:Christiane L. Joost-Gaugier, Pythagoras and Renaissance Europe: Finding Heaven, Cambridge University Press, 2009.
[12]:Oskar, Kristeller Paul. Studies in Renaissance thought and letters. IV. Roma: Edizioni di Storia e letteratura, 1996: 565.
[13]:“Three Books on Life”. World Digital Library. 26 February 2014. Retrieved 1 March 2014.
[14]:Deitz, Luc; Kraye, Jill (1997). “Marsilio Ficino”. Cambridge Translations of Renaissance Philosophical Texts. pp. 147–155. doi:10.1017/CBO9780511803048.014. ISBN 9780511803048.
[15]:Daniel Pickering Walker (January 2000). Spiritual and Demonic Magic: From Ficino to Campanella. Penn State Press. p. 3. ISBN 0-271-02045-8.
[16]:Marsilio Ficino, Three Books on Life, translated by Carol V. Kaske and John R. Clark, Tempe AZ: The Renaissance Society of America, 2002. From the Apologia, p. 399. (The internal quote is from Acts 17:28.)
[17]:Marsilio Ficino. Biography and introduction to The Letters of Marsilio Ficino, Volume 1 Archived 22 July 2014 at the Wayback Machine 1975 Fellowship of the School of Economic Science, London. Retrieved 26 April 2014.
[18]:Kaske, Carol (2006). “Review: Marsilio Ficino. The Letters of Marsilio Ficino”. Renaissance Quarterly. 59 (3): 829. doi:10.1353/ren.2008.0389. JSTOR 10.1353/ren.2008.0389. S2CID 164146779 – via JSTOR. I find no evidence in his letters of the homosexuality of which some contemporaries and some scholars over the last fifty years have suspected him.

- -
- - - -
- - - - - -
-
- - - - -
- - 0% -
- - - - - - - - - - - - - - - - - - - - - - - - - diff --git "a/blog/\343\200\220WORK-development record\343\200\221EQCat.html" "b/blog/\343\200\220WORK-development record\343\200\221EQCat.html" deleted file mode 100644 index 1017879..0000000 --- "a/blog/\343\200\220WORK-development record\343\200\221EQCat.html" +++ /dev/null @@ -1,276 +0,0 @@ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -【WORK-development record】EQCat | 尚 Shawn's Personal Page - - - - - - - - - - - - - - -
- -
-
-
- - - - - -
- - - - - - - - -
- - - - - -
- -
- - - - - -
- -

【WORK-development record】EQCat -

- - - -
- - - -
-

Development record of this application.

- -

brief summary:

A simple chatbot UI framework for an AI application. With a image of chat history, the application provides proper reply messages to their bosses or colleges for the User, including a domain model for segmenting and extracting text from WeChat chat history.

-

With the help of backend programs, EQCat follows a structure using different methods to ask the User to provide more information about the situation and problems they face, and reply tem the potential messages they could use.

- -
- - - -
- - - - - -
-
- - - - -
- - 0% -
- - - - - - - - - - - - - - - - - - - - - - - - - diff --git "a/blog/\343\200\220WORK-intern experience\343\200\221BeijingAcademyOfScienceAndTechnology - \345\211\257\346\234\254.html" "b/blog/\343\200\220WORK-intern experience\343\200\221BeijingAcademyOfScienceAndTechnology - \345\211\257\346\234\254.html" deleted file mode 100644 index 8fa523f..0000000 --- "a/blog/\343\200\220WORK-intern experience\343\200\221BeijingAcademyOfScienceAndTechnology - \345\211\257\346\234\254.html" +++ /dev/null @@ -1,276 +0,0 @@ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -【WORK-intern experience】Beijing Academy of Science And Technology | 尚 Shawn's Personal Page - - - - - - - - - - - - - - -
- -
-
-
- - - - - -
- - - - - - - - -
- - - - - -
- -
- - - - - -
- -

【WORK-intern experience】Beijing Academy of Science And Technology -

- - - -
- - - -
-

summary of the works I’ve done in this intern experience.

- -

brief summary:

Institute of Urban Safety and Environmental Science, Beijing Academy of Science and Technology
Beijing, China, 07/2023-09/2023

-

I contributed to the EU’s “Horizon 2020” AQ-WATCH project(Chinese team), where I built pipelines for high-dimensional pollution emission grid dataset over 200GB, and replace former CNN+LSTM system by applying cutting-edge Spatio-Temporal deep learning models like different kinds of Transformers, which reduces time costs and enhances accuracy.

- -
- - - -
- - - - - -
-
- - - - -
- - 0% -
- - - - - - - - - - - - - - - - - - - - - - - - - diff --git a/categories/Translation/index.html b/categories/Translation/index.html index 17f4334..3384cc9 100644 --- a/categories/Translation/index.html +++ b/categories/Translation/index.html @@ -84,7 +84,7 @@ diff --git a/categories/areas/index.html b/categories/areas/index.html index 3c47800..e176910 100644 --- a/categories/areas/index.html +++ b/categories/areas/index.html @@ -84,7 +84,7 @@ diff --git a/categories/books/index.html b/categories/books/index.html index 43db52c..738ee6c 100644 --- a/categories/books/index.html +++ b/categories/books/index.html @@ -84,7 +84,7 @@ diff --git a/categories/docs/index.html b/categories/docs/index.html index 1c8efab..c869f0b 100644 --- a/categories/docs/index.html +++ b/categories/docs/index.html @@ -84,7 +84,7 @@ diff --git a/categories/index.html b/categories/index.html index 21fbdd9..1b22d91 100644 --- a/categories/index.html +++ b/categories/index.html @@ -87,7 +87,7 @@ diff --git a/categories/math/computer-science/index.html b/categories/math/computer-science/index.html index 8b72ede..09623f6 100644 --- a/categories/math/computer-science/index.html +++ b/categories/math/computer-science/index.html @@ -84,7 +84,7 @@ diff --git a/categories/math/index.html b/categories/math/index.html index edffbb1..b9cd4e2 100644 --- a/categories/math/index.html +++ b/categories/math/index.html @@ -84,7 +84,7 @@ diff --git a/categories/other/index.html b/categories/other/index.html index 5dec19f..6aa7f73 100644 --- a/categories/other/index.html +++ b/categories/other/index.html @@ -84,7 +84,7 @@ diff --git a/categories/papers/index.html b/categories/papers/index.html index 1b5cf4f..89b46d3 100644 --- a/categories/papers/index.html +++ b/categories/papers/index.html @@ -84,7 +84,7 @@ diff --git a/categories/philosophy/index.html b/categories/philosophy/index.html index ad44a58..5f5e8b9 100644 --- a/categories/philosophy/index.html +++ b/categories/philosophy/index.html @@ -84,7 +84,7 @@ diff --git a/categories/works/index.html b/categories/works/index.html index 60099c1..202d925 100644 --- a/categories/works/index.html +++ b/categories/works/index.html @@ -84,7 +84,7 @@ diff --git a/index.html b/index.html index 295d4f3..6b70a04 100644 --- a/index.html +++ b/index.html @@ -87,7 +87,7 @@ diff --git a/tags/DIY/index.html b/tags/DIY/index.html index f6b30aa..a7615a4 100644 --- a/tags/DIY/index.html +++ b/tags/DIY/index.html @@ -84,7 +84,7 @@ diff --git a/tags/Deep-Learning/index.html b/tags/Deep-Learning/index.html index 8adcea9..c22e1ab 100644 --- a/tags/Deep-Learning/index.html +++ b/tags/Deep-Learning/index.html @@ -84,7 +84,7 @@ diff --git a/tags/HomeLab/index.html b/tags/HomeLab/index.html index 9894897..648fa70 100644 --- a/tags/HomeLab/index.html +++ b/tags/HomeLab/index.html @@ -84,7 +84,7 @@ diff --git a/tags/NLP/index.html b/tags/NLP/index.html index 9412d81..1af530d 100644 --- a/tags/NLP/index.html +++ b/tags/NLP/index.html @@ -84,7 +84,7 @@ diff --git a/tags/causal-AI/index.html b/tags/causal-AI/index.html index aef7f59..90c6c3d 100644 --- a/tags/causal-AI/index.html +++ b/tags/causal-AI/index.html @@ -84,7 +84,7 @@ diff --git a/tags/causal/index.html b/tags/causal/index.html index d22afb8..56bb90d 100644 --- a/tags/causal/index.html +++ b/tags/causal/index.html @@ -84,7 +84,7 @@ diff --git a/tags/generative-AI/index.html b/tags/generative-AI/index.html index 01ce822..51581c0 100644 --- a/tags/generative-AI/index.html +++ b/tags/generative-AI/index.html @@ -84,7 +84,7 @@ diff --git a/tags/generative-models/index.html b/tags/generative-models/index.html index 02e9077..ffc96e5 100644 --- a/tags/generative-models/index.html +++ b/tags/generative-models/index.html @@ -84,7 +84,7 @@ diff --git a/tags/index.html b/tags/index.html index 555c22f..a8412cf 100644 --- a/tags/index.html +++ b/tags/index.html @@ -87,7 +87,7 @@ diff --git a/tags/linux/index.html b/tags/linux/index.html index 8ea22bf..e34c3d1 100644 --- a/tags/linux/index.html +++ b/tags/linux/index.html @@ -84,7 +84,7 @@ diff --git a/tags/literature/index.html b/tags/literature/index.html index 8b54419..6d511e9 100644 --- a/tags/literature/index.html +++ b/tags/literature/index.html @@ -84,7 +84,7 @@ diff --git a/tags/machine-learning/index.html b/tags/machine-learning/index.html index a571c03..8bed71c 100644 --- a/tags/machine-learning/index.html +++ b/tags/machine-learning/index.html @@ -84,7 +84,7 @@ diff --git a/tags/math/index.html b/tags/math/index.html index 9441418..8377a3b 100644 --- a/tags/math/index.html +++ b/tags/math/index.html @@ -84,7 +84,7 @@ diff --git a/tags/metaworks/index.html b/tags/metaworks/index.html index 9e83287..1de89eb 100644 --- a/tags/metaworks/index.html +++ b/tags/metaworks/index.html @@ -84,7 +84,7 @@ diff --git a/tags/novel/index.html b/tags/novel/index.html index 3a1a16d..c3eedc3 100644 --- a/tags/novel/index.html +++ b/tags/novel/index.html @@ -84,7 +84,7 @@ diff --git a/tags/play/index.html b/tags/play/index.html index 35fe275..7e65efc 100644 --- a/tags/play/index.html +++ b/tags/play/index.html @@ -84,7 +84,7 @@ diff --git a/tags/probabilistic/index.html b/tags/probabilistic/index.html index f565547..7044562 100644 --- a/tags/probabilistic/index.html +++ b/tags/probabilistic/index.html @@ -84,7 +84,7 @@ diff --git a/tags/summary/index.html b/tags/summary/index.html index 8f72dac..0f8f312 100644 --- a/tags/summary/index.html +++ b/tags/summary/index.html @@ -84,7 +84,7 @@ diff --git a/tags/the-Renaissance/index.html b/tags/the-Renaissance/index.html index 780b07b..99f8d73 100644 --- a/tags/the-Renaissance/index.html +++ b/tags/the-Renaissance/index.html @@ -84,7 +84,7 @@ diff --git a/tags/time-series/index.html b/tags/time-series/index.html index 77e9836..2c92343 100644 --- a/tags/time-series/index.html +++ b/tags/time-series/index.html @@ -84,7 +84,7 @@ diff --git a/tags/video/index.html b/tags/video/index.html index 450fa57..d28518b 100644 --- a/tags/video/index.html +++ b/tags/video/index.html @@ -84,7 +84,7 @@ diff --git a/tags/wiki/index.html b/tags/wiki/index.html index ae7fb2e..3211b63 100644 --- a/tags/wiki/index.html +++ b/tags/wiki/index.html @@ -84,7 +84,7 @@ diff --git "a/\343\200\220AREA-introduction\343\200\221Causal Science/index.html" "b/\343\200\220AREA-introduction\343\200\221Causal Science/index.html" index 1b0824a..89424cb 100644 --- "a/\343\200\220AREA-introduction\343\200\221Causal Science/index.html" +++ "b/\343\200\220AREA-introduction\343\200\221Causal Science/index.html" @@ -88,7 +88,7 @@ diff --git "a/\343\200\220AREA-introduction\343\200\221Generative models/index.html" "b/\343\200\220AREA-introduction\343\200\221Generative models/index.html" index 274cad3..2d5759a 100644 --- "a/\343\200\220AREA-introduction\343\200\221Generative models/index.html" +++ "b/\343\200\220AREA-introduction\343\200\221Generative models/index.html" @@ -88,7 +88,7 @@ diff --git "a/\343\200\220AREA-introduction\343\200\221LLM/index.html" "b/\343\200\220AREA-introduction\343\200\221LLM/index.html" index d809fbb..858473e 100644 --- "a/\343\200\220AREA-introduction\343\200\221LLM/index.html" +++ "b/\343\200\220AREA-introduction\343\200\221LLM/index.html" @@ -90,7 +90,7 @@ diff --git "a/\343\200\220AREA-introduction\343\200\221time series/index.html" "b/\343\200\220AREA-introduction\343\200\221time series/index.html" index b9cb07f..3ed6c83 100644 --- "a/\343\200\220AREA-introduction\343\200\221time series/index.html" +++ "b/\343\200\220AREA-introduction\343\200\221time series/index.html" @@ -87,7 +87,7 @@ diff --git "a/\343\200\220BOOK-reading note\343\200\221Shakespeare Speaks/index.html" "b/\343\200\220BOOK-reading note\343\200\221Shakespeare Speaks/index.html" index 88b91c2..577f30f 100644 --- "a/\343\200\220BOOK-reading note\343\200\221Shakespeare Speaks/index.html" +++ "b/\343\200\220BOOK-reading note\343\200\221Shakespeare Speaks/index.html" @@ -88,7 +88,7 @@ diff --git "a/\343\200\220BOOK-reading note\343\200\221The Hairy Ape/index.html" "b/\343\200\220BOOK-reading note\343\200\221The Hairy Ape/index.html" index 2bc4d73..9a03d4b 100644 --- "a/\343\200\220BOOK-reading note\343\200\221The Hairy Ape/index.html" +++ "b/\343\200\220BOOK-reading note\343\200\221The Hairy Ape/index.html" @@ -88,7 +88,7 @@ diff --git "a/\343\200\220BOOK-reading note\343\200\221The Three Penny Opera/index.html" "b/\343\200\220BOOK-reading note\343\200\221The Three Penny Opera/index.html" index f75bbd4..e934665 100644 --- "a/\343\200\220BOOK-reading note\343\200\221The Three Penny Opera/index.html" +++ "b/\343\200\220BOOK-reading note\343\200\221The Three Penny Opera/index.html" @@ -88,7 +88,7 @@ diff --git "a/\343\200\220DOCS-notes\343\200\221how to build a NAS/index.html" "b/\343\200\220DOCS-notes\343\200\221how to build a NAS/index.html" index 8cf0259..9f94116 100644 --- "a/\343\200\220DOCS-notes\343\200\221how to build a NAS/index.html" +++ "b/\343\200\220DOCS-notes\343\200\221how to build a NAS/index.html" @@ -88,7 +88,7 @@ diff --git "a/\343\200\220DOCS-notes\343\200\221ubuntu/index.html" "b/\343\200\220DOCS-notes\343\200\221ubuntu/index.html" index c878350..c4e6a14 100644 --- "a/\343\200\220DOCS-notes\343\200\221ubuntu/index.html" +++ "b/\343\200\220DOCS-notes\343\200\221ubuntu/index.html" @@ -87,7 +87,7 @@ diff --git "a/\343\200\220MATH-notes\343\200\221Differential Equation/index.html" "b/\343\200\220MATH-notes\343\200\221Differential Equation/index.html" index 42129a1..fe8e068 100644 --- "a/\343\200\220MATH-notes\343\200\221Differential Equation/index.html" +++ "b/\343\200\220MATH-notes\343\200\221Differential Equation/index.html" @@ -87,7 +87,7 @@ diff --git "a/\343\200\220MATH-notes\343\200\221N=NP problem/index.html" "b/\343\200\220MATH-notes\343\200\221N=NP problem/index.html" index c83db4f..3501965 100644 --- "a/\343\200\220MATH-notes\343\200\221N=NP problem/index.html" +++ "b/\343\200\220MATH-notes\343\200\221N=NP problem/index.html" @@ -87,7 +87,7 @@ diff --git "a/\343\200\220OTHER-novel\343\200\221AI novel competition/index.html" "b/\343\200\220OTHER-novel\343\200\221AI novel competition/index.html" index 754c6c8..91bee41 100644 --- "a/\343\200\220OTHER-novel\343\200\221AI novel competition/index.html" +++ "b/\343\200\220OTHER-novel\343\200\221AI novel competition/index.html" @@ -94,7 +94,7 @@ diff --git "a/\343\200\220PAPER-introduction\343\200\221CausalVAE/index.html" "b/\343\200\220PAPER-introduction\343\200\221CausalVAE/index.html" index 0504e32..917be44 100644 --- "a/\343\200\220PAPER-introduction\343\200\221CausalVAE/index.html" +++ "b/\343\200\220PAPER-introduction\343\200\221CausalVAE/index.html" @@ -89,7 +89,7 @@ diff --git "a/\343\200\220PAPER-introduction\343\200\221DDPM/index.html" "b/\343\200\220PAPER-introduction\343\200\221DDPM/index.html" index 9cdb9cb..a6ec17d 100644 --- "a/\343\200\220PAPER-introduction\343\200\221DDPM/index.html" +++ "b/\343\200\220PAPER-introduction\343\200\221DDPM/index.html" @@ -87,7 +87,7 @@ diff --git "a/\343\200\220PAPER-introduction\343\200\221VQ-VAE/index.html" "b/\343\200\220PAPER-introduction\343\200\221VQ-VAE/index.html" index 5d92d2f..727fad2 100644 --- "a/\343\200\220PAPER-introduction\343\200\221VQ-VAE/index.html" +++ "b/\343\200\220PAPER-introduction\343\200\221VQ-VAE/index.html" @@ -87,7 +87,7 @@ diff --git "a/\343\200\220PAPER-introduction\343\200\221VideoGPT/index.html" "b/\343\200\220PAPER-introduction\343\200\221VideoGPT/index.html" index 199a6b1..0d2d99f 100644 --- "a/\343\200\220PAPER-introduction\343\200\221VideoGPT/index.html" +++ "b/\343\200\220PAPER-introduction\343\200\221VideoGPT/index.html" @@ -88,7 +88,7 @@ diff --git "a/\343\200\220PAPER-summary\343\200\221Papers read 23/index.html" "b/\343\200\220PAPER-summary\343\200\221Papers read 23/index.html" index 472b453..078f946 100644 --- "a/\343\200\220PAPER-summary\343\200\221Papers read 23/index.html" +++ "b/\343\200\220PAPER-summary\343\200\221Papers read 23/index.html" @@ -88,7 +88,7 @@ diff --git "a/\343\200\220PHIL-guide\343\200\221What Causality tells in real life/index.html" "b/\343\200\220PHIL-guide\343\200\221What Causality tells in real life/index.html" index 1c732a7..04d37fd 100644 --- "a/\343\200\220PHIL-guide\343\200\221What Causality tells in real life/index.html" +++ "b/\343\200\220PHIL-guide\343\200\221What Causality tells in real life/index.html" @@ -88,7 +88,7 @@ diff --git "a/\343\200\220TRANS-work\343\200\221Marsilio Ficino/index.html" "b/\343\200\220TRANS-work\343\200\221Marsilio Ficino/index.html" index a21f5de..4a211e5 100644 --- "a/\343\200\220TRANS-work\343\200\221Marsilio Ficino/index.html" +++ "b/\343\200\220TRANS-work\343\200\221Marsilio Ficino/index.html" @@ -90,7 +90,7 @@ diff --git "a/\343\200\220WORK-development record\343\200\221EQCat/index.html" "b/\343\200\220WORK-development record\343\200\221EQCat/index.html" index 1064061..cc16aff 100644 --- "a/\343\200\220WORK-development record\343\200\221EQCat/index.html" +++ "b/\343\200\220WORK-development record\343\200\221EQCat/index.html" @@ -87,7 +87,7 @@ diff --git "a/\343\200\220WORK-intern experience\343\200\221BeijingAcademyOfScienceAndTechnology - \345\211\257\346\234\254/index.html" "b/\343\200\220WORK-intern experience\343\200\221BeijingAcademyOfScienceAndTechnology - \345\211\257\346\234\254/index.html" index c1da613..ce62544 100644 --- "a/\343\200\220WORK-intern experience\343\200\221BeijingAcademyOfScienceAndTechnology - \345\211\257\346\234\254/index.html" +++ "b/\343\200\220WORK-intern experience\343\200\221BeijingAcademyOfScienceAndTechnology - \345\211\257\346\234\254/index.html" @@ -87,7 +87,7 @@