ایده های جدیدی که در هفته گذشته مقالات آنها منتشر شد:
🔸 Better Language Models of Code through Self-Improvement
🔸 Introducing Segment Anything: Working toward the first foundation model for image segmentation
🔸 Instruction Tuning with GPT-4
🔸 8 Things to Know about LLMs
🔸 A Survey of Large Language Models
🔸 Do the Rewards Justify the Means? Measuring Trade-Offs Between Rewards and Ethical Behavior in the MACHIAVELLI Benchmark
🔸 Baize: An Open-Source Chat Model with Parameter-Efficient Tuning on Self-Chat Data
🔸 Summary of ChatGPT/GPT-4 Research and Perspective Towards the Future of Large Language Models
🔸 Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling
🔸 SegGPT: Segmenting Everything In Context
#مقاله #ایده_جذاب
به کانال ما ملحق شین👇👇
✅ @AI_DeepMind
🔸 Better Language Models of Code through Self-Improvement
🔸 Introducing Segment Anything: Working toward the first foundation model for image segmentation
🔸 Instruction Tuning with GPT-4
🔸 8 Things to Know about LLMs
🔸 A Survey of Large Language Models
🔸 Do the Rewards Justify the Means? Measuring Trade-Offs Between Rewards and Ethical Behavior in the MACHIAVELLI Benchmark
🔸 Baize: An Open-Source Chat Model with Parameter-Efficient Tuning on Self-Chat Data
🔸 Summary of ChatGPT/GPT-4 Research and Perspective Towards the Future of Large Language Models
🔸 Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling
🔸 SegGPT: Segmenting Everything In Context
#مقاله #ایده_جذاب
به کانال ما ملحق شین👇👇
✅ @AI_DeepMind
👍3👎3👏1
Natural Language Reasoning, A Survey
An overview of natural language reasoning in NLP. Contains lots of discussions around language models, capabilities, limitations, and open questions.
arxiv.org/abs/2303.14725
#مقاله
به کانال ما ملحق شین👇👇
✅ @AI_DeepMind
An overview of natural language reasoning in NLP. Contains lots of discussions around language models, capabilities, limitations, and open questions.
arxiv.org/abs/2303.14725
#مقاله
به کانال ما ملحق شین👇👇
✅ @AI_DeepMind
Generative Agents: Interactive Simulacra of Human Behavior
arxiv.org/abs/2304.03442
#مقاله
✅ @AI_DeepMind
arxiv.org/abs/2304.03442
#مقاله
✅ @AI_DeepMind
برای درک درست transformers/LLMs این سه مقاله و این کتاب پیشنهاد میشود.
1- Overviews of transformer architectures and algorithms.
arxiv.org/pdf/2207.09238
2- The Survey of Large Language Models overviews the history of LMs and describes the ideas behind each model.
arxiv.org/abs/2303.18223
3- Transformer models
arxiv.org/pdf/2302.07730
4- Mastering Transformers: Build State-of-the-art Models from Scratch with Advanced Natural Language Processing Techniques
#مقاله #کتاب #ایده_جذاب
✅ @AI_DeepMind
1- Overviews of transformer architectures and algorithms.
arxiv.org/pdf/2207.09238
2- The Survey of Large Language Models overviews the history of LMs and describes the ideas behind each model.
arxiv.org/abs/2303.18223
3- Transformer models
arxiv.org/pdf/2302.07730
4- Mastering Transformers: Build State-of-the-art Models from Scratch with Advanced Natural Language Processing Techniques
#مقاله #کتاب #ایده_جذاب
✅ @AI_DeepMind
👍2❤1
Grounded-Segment-Anything: Marrying Grounding DINO with Segment Anything & Stable Diffusion & BLIP - Automatically Detect , Segment and Generate Anything with Image and Text Inputs
https://github.com/IDEA-Research/Grounded-Segment-Anything
Demo:
https://github.com/IDEA-Research/Grounded-Segment-Anything/blob/main/gradio_app.py
#مقاله
✅ @AI_DeepMind
https://github.com/IDEA-Research/Grounded-Segment-Anything
Demo:
https://github.com/IDEA-Research/Grounded-Segment-Anything/blob/main/gradio_app.py
#مقاله
✅ @AI_DeepMind
InstantBooth: Personalized Text-to-Image Generation without Test-Time Finetuning
model can instantly generate personalized images with only a single forward pass
Paper:
arxiv.org/abs/2304.03411
project page:
jshi31.github.io/InstantBooth/
#مقاله
🔸 مطالب بیشتر 👇👇
✅ @AI_DeepMind
model can instantly generate personalized images with only a single forward pass
Paper:
arxiv.org/abs/2304.03411
project page:
jshi31.github.io/InstantBooth/
#مقاله
🔸 مطالب بیشتر 👇👇
✅ @AI_DeepMind
تیم هوش مصنوعی دانشگاه MIT الگوریتم Robust MADER که قادر است مسیرهای بدون برخوردی را برای هواپیماهای بدون سرنشین ایجاد کند در مقاله زیر چگونگی حل آنرا شرح داده است، حتی اگر زمانی که ارتباطات بین عوامل هوایی تاخیر وجود داشته باشد. این سیستم یک برنامه‌ریز مسیر ناهمزمان، غیرمتمرکز و چند‌عاملی است، به این معنی که هر پهپاد مسیر خود را فرموله می‌کند و سپس با پهپادهای نزدیک بررسی کرده تا مطمئن شود با هیچ‌یک از آنها برخورد نمی‌کند.
https://arxiv.org/abs/2209.13667
#مقاله
🔸 مطالب بیشتر 👇👇
✅ @AI_DeepMind
https://arxiv.org/abs/2209.13667
#مقاله
🔸 مطالب بیشتر 👇👇
✅ @AI_DeepMind
👍2
اینطوریاس دیگه درساتون رو بخونید و تخصصی چیزیو یادبگیرید که هوش مصنوعی ازتون نگیردتش.
ChatGPT Can Convert Natural Language Instructions Into Executable Robot Actions
-Outputs sequence of actions
-Represents operating environment formally
-Infers updated state
-Users can modify LLM prompt using natural language for robust robot operations
arxiv.org/abs/2304.03893
#مقاله #پردازش_زبان_طبیعی
🔸 مطالب بیشتر 👇👇
✅ @AI_DeepMind
ChatGPT Can Convert Natural Language Instructions Into Executable Robot Actions
-Outputs sequence of actions
-Represents operating environment formally
-Infers updated state
-Users can modify LLM prompt using natural language for robust robot operations
arxiv.org/abs/2304.03893
#مقاله #پردازش_زبان_طبیعی
🔸 مطالب بیشتر 👇👇
✅ @AI_DeepMind