-
Notifications
You must be signed in to change notification settings - Fork 0
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
Showing
7 changed files
with
126 additions
and
2 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,36 @@ | ||
--- | ||
title: "Beyond Object Recognition: A New Benchmark towards Object Concept Learning" | ||
tags: | ||
- 科研 | ||
categories: | ||
date: 2024-12-24T21:04:22+08:00 | ||
modify: 2024-12-24T21:04:22+08:00 | ||
dir: | ||
share: false | ||
cdate: 2024-12-24 | ||
mdate: 2024-12-24 | ||
--- | ||
|
||
# Beyond Object Recognition: A New Benchmark towards Object Concept Learning | ||
|
||
> [!info]+ 文章信息 | ||
> - 文章题目: Beyond Object Recognition: A New Benchmark towards Object Concept Learning | ||
> - 作者:[Yong-Lu Li](https://arxiv.org/search/cs?searchtype=author&query=Li,+Y), [Yue Xu](https://arxiv.org/search/cs?searchtype=author&query=Xu,+Y), [Xinyu Xu](https://arxiv.org/search/cs?searchtype=author&query=Xu,+X), [Xiaohan Mao](https://arxiv.org/search/cs?searchtype=author&query=Mao,+X), [Yuan Yao](https://arxiv.org/search/cs?searchtype=author&query=Yao,+Y), [Siqi Liu](https://arxiv.org/search/cs?searchtype=author&query=Liu,+S), [Cewu Lu](https://arxiv.org/search/cs?searchtype=author&query=Lu,+C) | ||
> - arXiv:[\[2212.02710\] Beyond Object Recognition: A New Benchmark towards Object Concept Learning](https://arxiv.org/abs/2212.02710) | ||
> - 代码:[GitHub - silicx/ObjectConceptLearning: This the official repository of OCL (ICCV 2023).](https://github.com/silicx/ObjectConceptLearning) | ||
> [!abstract]- Abstract | ||
> Understanding objects is a central building block of AI, especially for embodied AI. Even though object recognition excels with deep learning, current machines struggle to learn higher-level knowledge, e.g., what attributes an object has, and what we can do with it. Here, we propose a challenging Object Concept Learning (OCL) task to push the envelope of object understanding. It requires machines to reason out affordances and simultaneously give the reason: what attributes make an object possess these affordances. To support OCL, we build a densely annotated knowledge base including extensive annotations for three levels of object concept (category, attribute, affordance), and the clear causal relations of three levels. By analyzing the causal structure of OCL, we present a baseline, Object Concept Reasoning Network (OCRN). It leverages concept instantiation and causal intervention to infer the three levels. In experiments, OCRN effectively infers the object knowledge while following the causalities well. Our data and code are available at https://mvig-rhos.com/ocl. | ||
> [!note]- Conclusion | ||
> In this work, we introduce object concept learning (OCL) expecting machines to infer affordances and explain what attributes enable an object to possess them. Accordingly, we build an extensive dataset and present OCRN based on casual intervention and instantiation. OCRN achieves decent performance and follows the causalities well. However, OCL remains challenging and would inspire a line of studies on reasoning-based object understanding. | ||
## Model | ||
|
||
- 通过 attribute 来预测 affordance,并推断是那些 attribute 起到了主要作用。 | ||
|
||
> [!example]- | ||
> ![image.png](https://raw.githubusercontent.com/WncFht/picture/main/20241225134159699.png) | ||
- 通过 casual intervetion 来避免 object category bias | ||
- model implement 看的还是有点模糊,以后再来补充 |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
|
@@ -11,7 +11,7 @@ cdate: 2024-12-02 | |
mdate: 2024-12-08 | ||
--- | ||
|
||
<!-- # 2024-W49-12 --> | ||
# 2024-W49-12 | ||
|
||
## PLAN | ||
|
||
|
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
|
@@ -4,7 +4,7 @@ tags: | |
- 周记 | ||
--- | ||
|
||
<!-- # 2024-W50-12 --> | ||
# 2024-W50-12 | ||
|
||
## Review | ||
|
||
|
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,64 @@ | ||
--- | ||
title: 2024-W52-12 | ||
tags: | ||
- 周记 | ||
categories: dairy | ||
date: 2024-12-23T12:06:20+08:00 | ||
modify: 2024-12-23T12:06:20+08:00 | ||
dir: dairy | ||
share: false | ||
cdate: 2024-12-23 | ||
mdate: 2024-12-23 | ||
--- | ||
|
||
# 2024-W52-12 | ||
|
||
## Review | ||
|
||
### Completed | ||
|
||
- [x] 完成推理实验 | ||
- 终于搞完 cs231n,然后做了个[小任务](https://github.com/WncFht/OCL_lab),就开始帮学长干活了。 | ||
- [x] 学会使用 [[../../整理好的/zotero 使用指南|zotero 使用指南]] | ||
- [x] 联系高中老师 | ||
- 把之前高中学习经验的东西从博客中移出了,另外建了一个[网站](https://wncfht.github.io/SurviveTYGXManual/)。 | ||
- 顺便邀请了一些高中同学分享分享经验 | ||
- 之后大概会返校宣讲,分享一些学习经验 / 三一啊啥的。🤔,~~我也成学长了~~。 | ||
- [x] 读论文,建立了一些基本概念 | ||
- 写了一些笔记,比如 | ||
- [[../../科研/ULIP-2|ULIP-2]] | ||
- [[CLIP]] | ||
- [[OCRN]] | ||
- [[SigLIP]] | ||
- 看了一下李沐读论文的系列 | ||
- 包括 Transformer, CLIP, CLIP 之后的工作串讲上下,多模态串讲上下 | ||
- 看了一部分的 GAMES 003,讲的很不错 | ||
- [x] 稍微运营了一下交流群 | ||
- [x] 修复了一下博客 | ||
- 现在博客可以完整地支持 obsidian 的双链和 callout 了 | ||
|
||
### Uncompleted | ||
|
||
- [ ] 学完 CS61C | ||
- 忙不过来,甚至都还没捡起来 | ||
- [ ] 规划 6.s081 学习 | ||
- 忙不过来捏,而且感觉最近的重心在科研上,打算把领域内的论文看的差不多了,再来打系统方面的基础。 | ||
|
||
## Next Week Plan | ||
|
||
- [ ] 学习 CS61C,目前是学到第 6 个 lab,希望一个星期能学完剩下的。 | ||
- [ ] 看完李沐读论文的系列,然后下载论文做笔记,串联自己的知识图谱。 | ||
- [ ] 着手实验室工作。 | ||
|
||
## THOUGHTS | ||
|
||
- 感觉自己的大一上还是处在东玩玩西玩玩的状态,弄了不少跟以后方向不太有关的东西。趁寒假有空,早点想清楚想干什么吧。 | ||
- 欸,还是引用一段话来的贴切 | ||
|
||
> [!quote]+ | ||
> 其实归根结底,问题还是大多数人(包括我自己),最初根本想不清楚自己最后要干啥,做了好多无意义的事情。比如选一些无意义但难的课,(比如数分系列,如果不是真心喜欢数学千万别选),大创,学一些非 cs 的选修课,给一些水实验室打工等。刚上大学,都觉得自己有无穷的可能性,不屑于在写代码这一条路上吊死。诚然,有些人啥都搞还啥都能搞好,但是作为一个普通人,还是要认清自己能力的范围和最后的目标。看别人去社团,自己也想去凑凑,看别人做科研,自己也想搞搞;可是这样搞来搞去,发现曾经不屑于干的事情,现在连干的资格都没有,这时再后悔已经为时已晚。将来发展的方向还是越早想明白越好。 | ||
- 我想我自己绝对称不上努力,甚至很难说在追逐喜欢的东西,只不过感觉‘欸,干这个也还不错,那就这样吧’。最近确实认识到很多优秀的学长,他们要么是去 MSRA 实习,然后在科研的道路上越走越远;要么是一手抓后端,一手抓系统,去了量化赚米。 | ||
|
||
> [!failure] | ||
> 也许,我真的是缺乏 all in 的勇气吧。 |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,22 @@ | ||
--- | ||
title: zotero 使用指南 | ||
tags: | ||
- Zotero | ||
- Tools | ||
categories: | ||
date: 2024-12-23T12:16:42+08:00 | ||
modify: 2024-12-23T12:16:42+08:00 | ||
dir: | ||
share: false | ||
cdate: 2024-12-23 | ||
mdate: 2024-12-23 | ||
--- | ||
|
||
# zotero 使用指南 | ||
|
||
- [zotero 使用指南](https://blog.csdn.net/l903445981/article/details/119960684) | ||
- [Zotero 中文社区 \| Zotero 中文维护小组](https://zotero-chinese.com/) | ||
- [ZOTERO 与 Obsidian 笔记插入和联动(含美化的笔记模板)](https://zhuanlan.zhihu.com/p/677461923) | ||
- [ZOTERO 于 Obsidian 联动方案](https://zhuanlan.zhihu.com/p/651144180) | ||
- 看了一圈下来,感觉现有方案都不太符合心意,于是不折腾了。不如让 zotero 和 obsidian 分开,真要做笔记的时候肯定是直接写出来的,没必要交织在一起了。 | ||
- 安装了一个插件 translate to pdf 以外就没有搞什么了。 |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters