Skip to main content

🏒 Microsoft Research

Ensembling Large Language Models with Process Reward-Guided Tree Search for Better Complex Reasoning
·4085 words·20 mins· loading · loading
AI Generated πŸ€— Daily Papers Natural Language Processing Large Language Models 🏒 Microsoft Research
λŒ€κ·œλͺ¨ μ–Έμ–΄ λͺ¨λΈλ“€μ˜ 앙상블을 톡해 λ³΅μž‘ν•œ μΆ”λ‘  문제λ₯Ό λ”μš± 효과적으둜 ν•΄κ²°ν•˜λŠ” μƒˆλ‘œμš΄ ν”„λ ˆμž„μ›Œν¬, LE-MCTSλ₯Ό μ œμ•ˆν•©λ‹ˆλ‹€!
MixLLM: LLM Quantization with Global Mixed-precision between Output-features and Highly-efficient System Design
·2237 words·11 mins· loading · loading
AI Generated πŸ€— Daily Papers Natural Language Processing Large Language Models 🏒 Microsoft Research
MixLLM: 좜λ ₯ νŠΉμ§• κ°„μ˜ μ „μ—­ ν˜Όν•© 정밀도 μ–‘μžν™”μ™€ 고효율 μ‹œμŠ€ν…œ 섀계λ₯Ό 톡해 LLM의 정확도와 νš¨μœ¨μ„±μ„ λ™μ‹œμ— ν–₯μƒμ‹œν‚€λŠ” 획기적인 μ–‘μžν™” 방법
VidTok: A Versatile and Open-Source Video Tokenizer
·2469 words·12 mins· loading · loading
AI Generated πŸ€— Daily Papers Computer Vision Video Understanding 🏒 Microsoft Research
VidTok: μ˜€ν”ˆμ†ŒμŠ€ κ³ μ„±λŠ₯ λΉ„λ””μ˜€ ν† ν¬λ‚˜μ΄μ €κ°€ 연속 및 이산 ν† ν°ν™”μ—μ„œ μ΅œμ²¨λ‹¨ μ„±λŠ₯을 λ‹¬μ„±ν•˜λ©°, 효율적인 ν•™μŠ΅ μ „λž΅κ³Ό ν˜μ‹ μ μΈ μ–‘μžν™” 기법을 톡해 μ˜μƒ 생성 및 이해 연ꡬ에 μƒˆλ‘œμš΄ κ°€λŠ₯성을 μ—΄μ—ˆμŠ΅λ‹ˆλ‹€.
TraceVLA: Visual Trace Prompting Enhances Spatial-Temporal Awareness for Generalist Robotic Policies
·2744 words·13 mins· loading · loading
AI Generated πŸ€— Daily Papers AI Applications Robotics 🏒 Microsoft Research
TraceVLA: 과거의 μ›€μ§μž„μ„ μ‹œκ°μ μœΌλ‘œ λ³΄μ—¬μ€ŒμœΌλ‘œμ¨ λ‘œλ΄‡μ˜ μ‹œκ³΅κ°„μ  인식을 ν–₯μƒμ‹œν‚΅λ‹ˆλ‹€.
Phi-4 Technical Report
·2236 words·11 mins· loading · loading
AI Generated πŸ€— Daily Papers Natural Language Processing Large Language Models 🏒 Microsoft Research
Phi-4: 140μ–΅ λ§€κ°œλ³€μˆ˜ μ–Έμ–΄ λͺ¨λΈμ€ 데이터 ν’ˆμ§ˆμ— 쀑점을 λ‘” ν›ˆλ ¨ λ ˆμ‹œν”Όλ‘œ κ°œλ°œλ˜μ–΄ μΆ”λ‘  λŠ₯λ ₯을 λŒ€ν­ ν–₯μƒμ‹œμΌ°μŠ΅λ‹ˆλ‹€.