bequiet-log
๐Ÿท๏ธ Tags
๐Ÿ’ป Profile
profile_image
Yongjun Cho
Machine Learning Researcher
Let's make Synergy together
๐Ÿ”Ž Search
๐Ÿ“‚ All Posts
๐Ÿค– Embodied AI
[Paper Review] OpenVLA: An Open-Source Vision-Language-Action Model

[Paper Review] OpenVLA: An Open-Source Vision-Language-Action Model

Aug 19, 2024

1. Yet, widespread adoption of VLAs for robotics has been challenging as 1) existing VLAs are largely closed and inaccessible to the public, and 2) prior work fails to explore methods for efficiently fine-tuning VLAs for new tasks, a key component for adoption. 2. Addressing these challenges, we introduce OpenVLA, a 7B-parameter open-source VLA trained on a diverse collection of 970k real-world robot demonstrations. OpenVLA builds on a Llama 2 language model combined with a visual encoder that fuses pretrained features from DINOv2 and SigLIP. 3. We further show that we can effectively fine-tune OpenVLA for new settings, with especially strong generalization results in multi-task environments involving multiple objects and strong language grounding abilities, and outperform expressive from-scratch imitation learning methods such as Diffusion Policy by 20.4%

Embodied AI
ENG
Blog
๐Ÿค– Embodied AI
[Paper Review] Mobility VLA: Multimodal Instruction Navigation with Long-Context VLMs and Topological Graphs

[Paper Review] Mobility VLA: Multimodal Instruction Navigation with Long-Context VLMs and Topological Graphs

Jul 12, 2024

โ€ข An elusive goal in navigation research is to build an intelligent agent that can understand multimodal instructions including natural language and image, and perform useful navigation โ€ข To achieve this, we study a widely useful category of navigation tasks we call Multimodal Instruction Navigation with demonstration Tours (MINT), in which the environment prior is provided through a previously recorded demonstration video. โ€ข We evaluated Mobility VLA in a 836$m^2$ real world environment and show that Mobility VLA has a high end-to-end success rates on previously unsolved multimodal instructions such as โ€œWhere should I return this?โ€ while holding a plastic bin.

Blog
ENG
Vison Language Model
๐Ÿค– Embodied AI
(2) An Introduction to Vision-Language Modeling: A Guide to VLM Training

(2) An Introduction to Vision-Language Modeling: A Guide to VLM Training

Jun 4, 2024

The document provides an overview of vision-language modeling (VLM) training strategies, discussing when to use contrastive models like CLIP, masking techniques, generative models, and pretrained backbones. It emphasizes the importance of grounding and alignment in VLMs, detailing methods such as instruction tuning and reinforcement learning from human feedback (RLHF). Additionally, it highlights advancements in models like LLaVA and its variants, which incorporate multimodal instruction tuning and improve performance on various benchmarks. Finally, it addresses parameter-efficient fine-tuning methods to adapt large-scale models for specific tasks while managing computational costs.

Vison Language Model
Blog
ENG
๐Ÿค– Embodied AI
(1) An Introduction to Vision-Language Modeling: The Families of VLMs

(1) An Introduction to Vision-Language Modeling: The Families of VLMs

Jun 3, 2024

The document discusses Vision-Language Models (VLMs), highlighting their role in solving rate-distortion problems by optimizing predictive information and constraining conditional densities. It covers various approaches, including generative-based VLMs that generate text and images, and examples like CoCa and CM3Leon which utilize multimodal generative techniques. The document also explores the use of pretrained backbones in VLMs, emphasizing models like MiniGPT and BLIP2 that efficiently integrate visual and textual data for various tasks, showcasing advancements in multimodal understanding and generation capabilities.

Blog
ENG
Vison Language Model
KOR
๐Ÿค– Embodied AI
[๋…ผ๋ฌธ๋ฆฌ๋ทฐ] Driving Everywhere with Large Language Model Policy Adaptation

[๋…ผ๋ฌธ๋ฆฌ๋ทฐ] Driving Everywhere with Large Language Model Policy Adaptation

May 24, 2024

โ€ข Adapting driving behavior to new environments, customs, and laws is a long-standing problem in autonomous driving, precluding the widespread deployment of autonomous vehicles (AVs) โ€ข LLaDA achieves this by leveraging the impressive zero-shot generalizability of large language models (LLMs) in interpreting the traffic rules in the local driver handbook. โ€ข We also demonstrate LLaDAโ€™s ability to adapt AV motion planning policies in real-world datasets; LLaDA outperforms baseline planning approaches on all our metrics.

KOR
ENG
Large Language Model
Blog
๐Ÿค– Embodied AI
(3) [Paper Review] Fine-Tuning Large Vision-Language Models as Decision-Making Agents via Reinforcement Learning: Embodied AI

(3) [Paper Review] Fine-Tuning Large Vision-Language Models as Decision-Making Agents via Reinforcement Learning: Embodied AI

May 21, 2024

Today, I will review a new paper that was released yesterday. This research comes from Sergey Levineโ€™s team, a prominent figure in the AI and RL domains. They propose fine-tuning Vision-Language Models (VLM) with Reinforcement Learning (RL) to enhance performance in optimal decision-making tasks within multi-step interactive environments. The paper presents a simple approach that outperforms both GPT-4 and Gemini. This research is similar to my own ideas for solving challenges in embodied AI. Therefore, I will review this paper and organize its key concepts.

ENG
Blog
Vison Language Model
๐Ÿค– Embodied AI
[๋…ผ๋ฌธ๋ฆฌ๋ทฐ] Scaling Instructable Agents Across Many Simulated Worlds

[๋…ผ๋ฌธ๋ฆฌ๋ทฐ] Scaling Instructable Agents Across Many Simulated Worlds

May 20, 2024

1. ์ž„์˜์˜ ์–ธ์–ด ์ง€์‹œ๋ฅผ ๋”ฐ๋ฅผ ์ˆ˜ ์žˆ๋„๋ก ๊ตฌ์ฒด์  ํ–‰๋™์— ๋งž์ถ”์–ด ๋ณต์žกํ•œ ์ž‘์—…์„ ์ˆ˜ํ–‰ํ•˜๋„๋ก ํ•œ๋‹ค. 2. ์‹œ๋ฎฌ๋ ˆ์ด์…˜๋œ 3D ํ™˜๊ฒฝ์—์„œ ์ธ๊ฐ„์ด ํ•  ์ˆ˜ ์žˆ๋Š” ๋ชจ๋“  ๊ฒƒ์„ ์ˆ˜ํ–‰ํ•  ์ˆ˜ ์žˆ๋Š” Scalable, Instructable, Multiworld agent๋ฅผ ํ•™์Šตํ•  ๊ฒƒ์ด๋‹ค. language + observation โ†’ keyboard-and-mouse 3. ๋™๊ธฐ์™€ ๋ชฉํ‘œ, ์ดˆ๊ธฐ ์ง„ํ–‰์ƒํ™ฉ, ์—ฌ๋Ÿฌ ์—ฐ๊ตฌํ™˜๊ฒฝ๊ณผ ์ƒ์—…์šฉ ๋น„๋””์˜ค ๊ฒŒ์ž„์—์„œ์˜ ์˜ˆ๋น„์ ์ธ ๊ฒฐ๊ณผ๋ฅผ ์„ค๋ช…ํ•œ๋‹ค.

Blog
KOR
Embodied AI
๐Ÿ“ฆ 3D Generation
(3) Align Your Gaussians: Text-to-4D with Dynamic 3D Gaussians and Composed Diffusion Models : 3D Generation

(3) Align Your Gaussians: Text-to-4D with Dynamic 3D Gaussians and Composed Diffusion Models : 3D Generation

May 14, 2024

์ด๋ฒˆ์—๋Š” 4D generation (3D generation + motion)์— ๋Œ€ํ•ด ๋ฆฌ๋ทฐํ•ด๋ณด๋„๋ก ํ•˜๊ฒ ๋‹ค. ์ด ์—ฐ๊ตฌ๋Š” Nvidia์—์„œ ๋ฐœํ‘œํ•œ ๋…ผ๋ฌธ์œผ๋กœ ์—ฌ๋Ÿฌ ๋ชจ๋ธ์„ ์‚ฌ์šฉํ•˜์—ฌ 4D generation์„ ์ง„ํ–‰ํ•˜๋Š” ๊ฒƒ์„ ๋ชฉํ‘œ๋กœํ•˜์˜€๋‹ค. ์•„์ง์€ ๋ฐœ์ „ํ•  ๊ฒƒ์ด ๋งŽ์•„๋ณด์ด์ง€๋งŒ ๊ทธ๋ž˜๋„ ์ƒˆ๋กœ์šด ์—ฐ๊ตฌ ๋ฐฉํ–ฅ์œผ๋กœ์จ 4D๊ฐ€ ์ฃผ๋ชฉ๋ฐ›๊ณ  ์žˆ๊ณ  ์—ฐ๊ตฌ๋ฅผ ํ•˜๊ธฐ์—๋Š” ์ตœ์ ์˜ ์ฃผ์ œ๋ผ๊ณ  ์ƒ๊ฐํ•œ๋‹ค.

3D Generation
KOR
Diffusion
Blog
๐Ÿค– Embodied AI
(2) [๋…ผ๋ฌธ๋ฆฌ๋ทฐ] RT-2, Vision-Language-Action Models Transfer Web Knowlege to Robotic Control: Embodied AI

(2) [๋…ผ๋ฌธ๋ฆฌ๋ทฐ] RT-2, Vision-Language-Action Models Transfer Web Knowlege to Robotic Control: Embodied AI

May 10, 2024

์ด๋ฒˆ์—๋Š” Q-transformer์— ์ด์–ด ๋”ฅ๋งˆ์ธ๋“œ์—์„œ ๊ณต๊ฐœํ•œ Embodied AI์ธ RT-2์— ๋Œ€ํ•œ ๋ฆฌ๋ทฐ๋ฅผ ์ง„ํ–‰ํ•ด๋ณด๋„๋ก ํ•˜๊ฒ ๋‹ค. ์ด์ „์— RT-1๊ณผ Q-transformer๊ฐ€ ๋กœ๋ด‡ ๋ฐ์ดํ„ฐ๋งŒ์„ ๊ฐ€์ง€๊ณ  Transformer๋ฅผ ํ•™์Šต์‹œ์ผœ Imitation learning์„ ์ง„ํ–‰ํ–ˆ๋‹ค๋ฉด ์ด๋ฒˆ์—๋Š” Internet Scale์—์„œ ํ•™์Šต๋œ Vision Language ๋ชจ๋ธ์„ ์‚ฌ์šฉํ•˜์—ฌ ๋กœ๋ด‡ ๋ฐ์ดํ„ฐ๋ฅผ ์ถ”๊ฐ€ํ•ด ๋”์šฑ Generalization ์„ฑ๋Šฅ์ด ๋›ฐ์–ด๋‚œ ๋ชจ๋ธ์„ ๊ฐœ๋ฐœํ•˜๋Š” ์—ฐ๊ตฌ์ด๋‹ค. GPT๋ฅผ ์จ๋ณด์•˜๋‹ค๋ฉด ์ด๋ฏธ ์ด๋ฏธ์ง€๋ฅผ ํ†ตํ•œ Reasoning์˜ ์ˆ˜์ค€์ด ๋†€๋ผ์šด ์ˆ˜์ค€์ด๊ณ , ์ด๋ฅผ ํ™œ์šฉํ•˜๋ฉด ์‹ค์ œ ๋กœ๋ด‡์„ ๋งŒ๋“ค ์ˆ˜ ์žˆ๋‹ค๋Š” ์ƒ์ƒ์„ ํ•  ์ˆ˜ ์žˆ์„ ๊ฒƒ์ด๋‹ค. ์ด ๋…ผ๋ฌธ์€ ๊ทธ ์ƒ์ƒ์„ ์ง์ ‘ ์‹คํ—˜์œผ๋กœ ์ฆ๋ช…ํ•˜๊ณ  ๊ฒ€์ฆํ•œ ๋…ผ๋ฌธ์ด๋‹ค. ์ด ๋…ผ๋ฌธ์„ ํ†ตํ•ด ๋” ์ข‹์€ ์„ฑ๋Šฅ์˜ Embodied AI๊ฐ€ ๊ฐœ๋ฐœ๋  ๊ฒƒ์ด๋ผ๋Š” ํ™•์‹ ์„ ๊ฐ€์ง€๊ฒŒ ๋˜์—ˆ๋‹ค.

KOR
Robotics
Vison Language Model
Blog
๐Ÿ“ฆ 3D Generation
(2) [๋…ผ๋ฌธ๋ฆฌ๋ทฐ] Zero123: 3D Generation

(2) [๋…ผ๋ฌธ๋ฆฌ๋ทฐ] Zero123: 3D Generation

May 9, 2024

์ด๋ฒˆ ํฌ์ŠคํŒ…์—์„œ๋Š” 3D generation ๋ชจ๋ธ ์ค‘์—์„œ Zero123๋ฅผ ๋ฆฌ๋ทฐํ•ด ๋ณผ ๊ฒƒ์ด๋‹ค. Zero123๋Š” diffusion model์„ ์นด๋ฉ”๋ผ ๊ฐ๋„์— ๋”ฐ๋ฅธ ์ด๋ฏธ์ง€๋ฅผ ์ƒ์„ฑํ•˜๋„๋ก finetuningํ•˜์—ฌ 3D generation์„ ์ง„ํ–‰ํ•œ๋‹ค๋Š” ๋งค์šฐ ๊ฐ„๋‹จํ•œ ์•„์ด๋””์–ด์—์„œ ์ถœ๋ฐœํ•œ ๋…ผ๋ฌธ์ด๋‹ค.

3D Generation
Blog
KOR
Diffusion
๐Ÿ“ฆ 3D Generation
(1) [๋…ผ๋ฌธ๋ฆฌ๋ทฐ] DreamFusion: Text-To-3D Using 2D Diffusion - 3D generation

(1) [๋…ผ๋ฌธ๋ฆฌ๋ทฐ] DreamFusion: Text-To-3D Using 2D Diffusion - 3D generation

Apr 29, 2024

์ด๋ฒˆ ํฌ์ŠคํŠธ์—์„œ๋Š” ํ˜„์žฌ ๋‹ค์–‘ํ•œ 3D generation ๋ชจ๋ธ์˜ ๊ธฐ์ดˆ๊ฐ€ ๋œ๋‹ค๊ณ  ๋ณผ ์ˆ˜ ์žˆ๋Š” DreamFusion์— ๋Œ€ํ•ด์„œ ๋ฆฌ๋ทฐํ•ด๋ณผ ๊ฒƒ์ด๋‹ค. ์ด ๋…ผ๋ฌธ์€ ๊ตฌ๊ธ€ ๋ฆฌ์„œ์น˜์™€ ๋ฒ„ํด๋ฆฌ์—์„œ ์ง„ํ–‰ํ•œ ์—ฐ๊ตฌ์ด๊ณ , 2D diffusion ๋ชจ๋ธ์„ ์‚ฌ์šฉํ•˜์—ฌ NeRF๋ฅผ ํ•™์Šต์‹œ์ผœ 3D generation ๋ชจ๋ธ์„ ๋งŒ๋“ค ์ˆ˜ ์žˆ๋‹ค๋Š” ์‚ฌ์‹ค๋กœ ์ฃผ๋ชฉ๋ฐ›์•˜๋‹ค. ํ˜„์žฌ๋Š” video prior, 2D์™€ 3D๋ฅผ ๊ฒฐํ•ฉํ•œ prior ๋“ฑ์„ ์ด์šฉํ•˜์—ฌ ๋‹ค์–‘ํ•œ ์—ฐ๊ตฌ๊ฐ€ ๋‚˜์˜ค๊ณ  ์žˆ๋‹ค. ์•ž์œผ๋กœ 3D generation ๋ชจ๋ธ๋“ค์— ๋Œ€ํ•ด ๋ฆฌ๋ทฐ๋ฅผ ์ง„ํ–‰ํ•˜๊ธฐ ์œ„ํ•ด ์•Œ์•„๋‘์–ด์•ผํ•˜๋Š” ๋…ผ๋ฌธ์ด๊ธฐ ๋•Œ๋ฌธ์— ์ž์„ธํ•œ ๋ฆฌ๋ทฐ๋ฅผ ์ง„ํ–‰ํ•ด๋ณด๋„๋ก ํ•  ๊ฒƒ์ด๋‹ค.

Diffusion
3D Generation
Blog
KOR
๐Ÿค– Embodied AI
(1) [๋…ผ๋ฌธ๋ฆฌ๋ทฐ] Q-transformer : Embodied AI

(1) [๋…ผ๋ฌธ๋ฆฌ๋ทฐ] Q-transformer : Embodied AI

Mar 4, 2024

์ด๋ฒˆ ํฌ์ŠคํŠธ์—์„œ๋Š” ์ง€๋‚œ๋ฒˆ ํฌ์ŠคํŠธ์—์„œ ์งง๊ฒŒ ์„ค๋ช…ํ–ˆ๋˜ Q-Transformer๋ผ๋Š” ๋…ผ๋ฌธ์— ๋Œ€ํ•ด ๋” ์ž์„ธํžˆ ์•Œ์•„๋ณผ ๊ฒƒ์ด๋‹ค. ์ด ํฌ์ŠคํŠธ๋ฅผ ์ดํ•ดํ•˜๊ธฐ ์œ„ํ•ด์„œ๋Š” ์ง€๋‚œ ๊ธ€ ์ค‘ offline RL ๋ถ€๋ถ„์„ ๋ฐ˜๋“œ์‹œ ์ฝ์–ด๋ณด๋Š” ๊ฒƒ์ด ์ข‹๋‹ค.

Reinforcement Learning
Blog
KOR
Robotics
๐Ÿ›ก๏ธReinforcement Learning
(2) Text-to-Image Diffusion Model, Alignment in Deep Learning : Comprehensive summary

(2) Text-to-Image Diffusion Model, Alignment in Deep Learning : Comprehensive summary

Feb 27, 2024

์ง€๋‚œ ํฌ์ŠคํŒ…์— ์ด์–ด ์ด๋ฒˆ์—๋Š” Image generation model์—์„œ์˜ alignment๋ฅผ ์‚ดํŽด๋ณด๋ ค๊ณ ํ•œ๋‹ค. ์ด ๋ถ„์•ผ๋Š” ํ˜„์žฌ ์น˜์—ดํ•œ ๊ฒฝ์Ÿ์ด ์ง„ํ–‰๋˜๊ณ  ์žˆ์–ด ๋งŽ์€ ๋…ผ๋ฌธ์ด ๋ฐœํ‘œ๋˜๊ณ  ์žˆ๋‹ค. ์ด ๊ธ€์—์„œ๋Š” ์ฒซ ์‹œ๋„์ธ Aligning Text-to-Image ๋…ผ๋ฌธ๋ถ€ํ„ฐ DPOK, Diffusion DPO๊นŒ์ง€ ์ž์„ธํ•˜๊ฒŒ ๋ฆฌ๋ทฐํ•ด๋ณด๊ณ ์ž ํ•œ๋‹ค. ๋‚˜๋จธ์ง€ ๋‹ค์–‘ํ•œ ์—ฐ๊ตฌ๋“ค์€ ์งง๊ฒŒ ์š”์•ฝํ•ด์„œ ์„ค๋ช…ํ•  ๊ฒƒ์ด๋‹ค.

Blog
KOR
Reinforcement Learning
Diffusion
๐Ÿ›ก๏ธReinforcement Learning
(1) RLHF LLM, Alignment in Deep Learning: Comprehensive Summary

(1) RLHF LLM, Alignment in Deep Learning: Comprehensive Summary

Feb 16, 2024

LLM๊ณผ Image generation ๋ชจ๋ธ์„ ํ†ตํ•ด ์ƒ์„ฑํ˜• ์ธ๊ณต์ง€๋Šฅ ๋ชจ๋ธ์— ๋Œ€ํ•œ ๊ด€์‹ฌ์€ ํญ๋ฐœ์ ์œผ๋กœ ์ฆ๊ฐ€ํ–ˆ๋‹ค. ์ด๋ฏธ ChatGPT์™€ ๋ฏธ๋“œ์ €๋‹ˆ์™€ ๊ฐ™์€ ์ธ๊ณต์ง€๋Šฅ ๋ชจ๋ธ์˜ ์˜ํ–ฅ๋ ฅ์€ ๊ฒฝ์ œ ์‚ฌํšŒ ์ „๋ฐ˜์˜ ๋ณ€ํ™”๋ฅผ ์ผ์œผํ‚ค๊ณ  ์žˆ๋‹ค. ํ•˜์ง€๋งŒ ์ƒ์„ฑ ๋ชจ๋ธ์„ ์„œ๋น„์Šค์— ์‚ฌ์šฉํ•˜๊ธฐ๊นŒ์ง€๋Š” ์—ฌ๋Ÿฌ ๋ฒˆ์˜ ํ•™์Šต ๊ณผ์ •์„ ๊ฑฐ์น˜๊ฒŒ ๋˜๋Š”๋ฐ, ์ด๋Š” ๋ฐ”๋กœ ์ƒ์„ฑํ˜• ๋ชจ๋ธ์ด ๊ฐ€์ง„ ํŠน์ง• ๋•Œ๋ฌธ์ด๋‹ค. ์ด ํ•™์Šต ๊ณผ์ •์—์„œ ํ•„์ˆ˜์ ์ธ Alignment์— ๋Œ€ํ•ด์„œ ์ •๋ฆฌํ•ด๋ณด์•˜๋‹ค.

Blog
KOR
Reinforcement Learning
Large Language Model
๐Ÿ˜Ž Daily
๋ฐ์ผ ์นด๋„ค๊ธฐ์˜ ์ธ๊ฐ„๊ด€๊ณ„๋ก ์„ ์ฝ๊ณ 

๋ฐ์ผ ์นด๋„ค๊ธฐ์˜ ์ธ๊ฐ„๊ด€๊ณ„๋ก ์„ ์ฝ๊ณ 

Jan 19, 2024

๋ฐ์ผ ์นด๋„ค๊ธฐ์˜ ์ธ๊ฐ„๊ด€๊ณ„๋ก ์„ ์ฝ๊ณ  ๋Š๋‚€์ ์„ ์ •๋ฆฌํ•ด๋ณด์•˜๋‹ค.

Blog
KOR
Book
๐Ÿ›ก๏ธReinforcement Learning
(4) ๊ฐ•ํ™”ํ•™์Šต ๊ฒŒ์ž„ ์ƒ์šฉํ™”  - [๋ชจ๋‘์˜ ์—ฐ๊ตฌ์†Œ] ์ค‘์š”ํ•œ ๊ฒƒ์€ ๊บพ์ด์ง€ ์•Š๋Š” RL ํ›„๊ธฐ

(4) ๊ฐ•ํ™”ํ•™์Šต ๊ฒŒ์ž„ ์ƒ์šฉํ™” - [๋ชจ๋‘์˜ ์—ฐ๊ตฌ์†Œ] ์ค‘์š”ํ•œ ๊ฒƒ์€ ๊บพ์ด์ง€ ์•Š๋Š” RL ํ›„๊ธฐ

Dec 19, 2023

์ด๋ฒˆ ๋ธ”๋กœ๊ทธ ๊ธ€์€ ๋ฒŒ์จ ๋งˆ์ง€๋ง‰ ์ฃผ์ œ์ด๋‹ค. ๋‘ ๊ฐœ์˜ ๋ฐœํ‘œ๋งŒ์ด ๋‚จ์•˜๋Š”๋ฐ ๋ชจ๋‘ ๊ฐ•ํ™”ํ•™์Šต์„ ๊ฒŒ์ž„์— ์ ์šฉํ•˜์—ฌ ์ƒ์šฉํ™”ํ•˜๋Š” ๋ฐฉ๋ฒ•์— ๋Œ€ํ•œ ๋ฐœํ‘œ์˜€๋‹ค.

Seminar
Blog
KOR
Reinforcement Learning
๐Ÿ˜Ž Daily
Striving for Visibility: My Journey to Get My Blog Indexed on Google

Striving for Visibility: My Journey to Get My Blog Indexed on Google

Dec 10, 2023

How to get indexing from google search console efficiently

Blog
ENG
๐Ÿ›ก๏ธReinforcement Learning
(3) Pretraining for inteligent robot - [๋ชจ๋‘์˜ ์—ฐ๊ตฌ์†Œ] ์ค‘์š”ํ•œ ๊ฒƒ์€ ๊บพ์ด์ง€ ์•Š๋Š” RL ํ›„๊ธฐ

(3) Pretraining for inteligent robot - [๋ชจ๋‘์˜ ์—ฐ๊ตฌ์†Œ] ์ค‘์š”ํ•œ ๊ฒƒ์€ ๊บพ์ด์ง€ ์•Š๋Š” RL ํ›„๊ธฐ

Nov 30, 2023

์ด๋ฒˆ ๋ฐœํ‘œ๋Š” reinforcement learning์—์„œ pretrain๊ณผ ๊ด€๋ จ๋œ ์ „์ฒด์ ์ธ ๋‚ด์šฉ์„ ์„ค๋ช…ํ•ด์ฃผ๋Š” ๊ฐ•์˜์˜€๋‹ค. ์‰ฌ์šด ์„ค๋ช…๊ณผ ํ•จ๊ป˜ ๋Œ€ํ‘œ์ ์ธ ์•Œ๊ณ ๋ฆฌ์ฆ˜์„ ์„ค๋ช…ํ•ด์ฃผ์–ด ๋งค์šฐ ๋„์›€์ด ๋งŽ์ด ๋˜์—ˆ๋˜ ๊ฒƒ ๊ฐ™๋‹ค.

Seminar
Blog
KOR
Reinforcement Learning
Robotics
๐Ÿ›ก๏ธReinforcement Learning
(2) Causal RL, Multi environment RL - [๋ชจ๋‘์˜ ์—ฐ๊ตฌ์†Œ] ์ค‘์š”ํ•œ ๊ฒƒ์€ ๊บพ์ด์ง€ ์•Š๋Š” RL ํ›„๊ธฐ

(2) Causal RL, Multi environment RL - [๋ชจ๋‘์˜ ์—ฐ๊ตฌ์†Œ] ์ค‘์š”ํ•œ ๊ฒƒ์€ ๊บพ์ด์ง€ ์•Š๋Š” RL ํ›„๊ธฐ

Nov 29, 2023

Causal RL๊ณผ multi environment RL์— ๋Œ€ํ•œ ๋ฐœํ‘œ๋ฅผ ์ •๋ฆฌํ•ด๋ณด์•˜๋‹ค.

Seminar
Blog
KOR
Reinforcement Learning
๐Ÿ›ก๏ธReinforcement Learning
(1) 2023 Reinforcement Learning Trend - [๋ชจ๋‘์˜ ์—ฐ๊ตฌ์†Œ] ์ค‘์š”ํ•œ ๊ฒƒ์€ ๊บพ์ด์ง€ ์•Š๋Š” RL ํ›„๊ธฐ

(1) 2023 Reinforcement Learning Trend - [๋ชจ๋‘์˜ ์—ฐ๊ตฌ์†Œ] ์ค‘์š”ํ•œ ๊ฒƒ์€ ๊บพ์ด์ง€ ์•Š๋Š” RL ํ›„๊ธฐ

Nov 25, 2023

๋ชจ๋‘์˜ ์—ฐ๊ตฌ์†Œ ๊ฐ•ํ™”ํ•™์Šต ์„ธ๋ฏธ๋‚˜์— ์ฐธ์„๋ณด์•˜๋‹ค.

Seminar
Blog
KOR
Reinforcement Learning
๐Ÿ’ป Profile
Yongjun Cho
Machine Learning Researcher
Let's make Synergy together
๐Ÿ’ฌ Contact
github
email
linkedin