🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Thrilled to share that I’ll be joining the College of Computing and Data Science at Nanyang Technological University (NTU) (NTU Singapore) as an Assistant Professor, starting in August 2025 🇸🇬🥳 I’ll continue my research on building trustworthy and continually adaptable multimodal AI,

EPiC Efficient Video Camera Control Learning with Precise Anchor-Video Guidance

🚨Check out our new video generation work EPiC! 🌟EPiC enables precise 3D camera trajectory control for both image-to-video generation and video-to-video generation! 💡Key highlights: ▶️ Efficient training within 16 GPU-hours ▶️ No need for paired video-camera trajectory data

Check out our new paper (EPiC) for video generation with camera-control 🔥 Here are the two highlights for easy and efficient training: ➡️The model can be trained directly on videos in the wild, without requiring extra camera trajectory annotations. ➡️With a novel

🚀Check out our new paper EPiC for video generation with efficient and precise 3D camera control! Just 16 GPU-hours (vs. 200+), with higher quality results! We innovate on both data & model-level: ✅Data: Visibility-based masking—no video-camera trajectory paired data needed

🚨 New paper alert! EPiC: Video Generation with Precise 3D Camera Control 🎬 Tackles both I2V & V2V tasks with: ▶️ Visibility-based masking—no need for video-camera trajectories ▶️ Lightweight ControlNet, guided by anchor video as structural prior Details in the thread !👇

Introducing EPiC - precise & efficient camera control for video generation! 📽️⚙️ Previous methods had drawbacks: ❌ Noisy anchor videos from point cloud estimates ❌ Expensive camera pose annotations ❌ 200+ GPU hours to train EPiC addresses this with: ✅ Visibility-based

Big congratulations to Prof. Jaehong Yoon on his new role at NTU Singapore! 🎉 He was an amazing mentor and advisor, always thoughtful, supportive, and full of sharp ideas. If you're thinking about applying to a PhD in multimodal AI etc, definitely reach out to him! Looking forward to

Excited to share that I’ll be joining UNC Computer Science and UNC NLP as a Postdoctoral Research Associate, working with the incredible Mohit Bansal! Can’t wait to collaborate with the amazing students and faculty there! 🎉 A huge thank you to my supervisor Reza Haffari, my colleagues at

Excited to share Video-Skill-CoT🎬🛠️– a new framework for domain-adaptive video reasoning with skill-aware Chain-of-Thought (CoT) supervision! ⚡️Key Highlights: ➡️ Automatically extracts domain-specific reasoning skills from questions and organizes them into a unified taxonomy,

Introducing Video-Skill-CoT 📽️ , a new framework for domain-adaptive video understanding with skill-specific chain-of-thought reasoning! ✅ Automatically discovers reasoning skills from video data ✅ Trains skill-specific expert modules with skill-specific CoT rationales ✅

🚨 New Release: Video-Skill-CoT! Domain-Adaptive, Skill-Based Video Reasoning💡 ✅ Automatically extracts domain-specific reasoning skills ✅ Generates tailored, skill-based CoT rationales ✅ Trains with skill-specific experts for stronger domain adaptation 🚀 Outperforms

🔥High-Quality Video Generation Accelerator🔥 ⚡️Dual-Expert Consistency Model (#DCM)⚡️ brings 10× speedup to video gen models (from 1.3B to 13B) with no quality drop * Now supports Hunyuan and Wan - Page: vchitect.github.io/DCM/ - Code: github.com/Vchitect/DCM

This paper proposes VIDEO-SKILL-COT to improve domain adaptation using skill-aware Chain-of-Thought supervision and expert learning modules. Methods 🔧: → The framework automatically constructs skill-based Chain-of-Thought annotations by extracting skills from questions,

How much do LLMs memorize? Meta and collaborators suggest that they can estimate model capacity by measuring memorization. "Models in the GPT family have an approximate capacity of 3.6 bits-per-parameter." Once capacity fills, generalization begins! More in my notes below:

Excited to share our new work, CLaMR! 🚀 We tackle multimodal content retrieval by jointly considering video, speech, OCR, and metadata. CLaMR learns to dynamically pick the right modality for your query, boosting retrieval by 25 nDCG@10 over single modality retrieval! 🧐

Big news! 🎉 I’m joining UNC-Chapel Hill as an Assistant Professor in Computer Science starting next year! Before that, I’ll be spending time OpenAI working on LLM privacy. UNC Computer Science UNC NLP

🚨 Check out our new paper: Frame Guidance — a powerful, training-free framework for video control in diffusion models! 🎬 ▶️ Supports multiple forms of control in video diffusion models, including keyframe-guided generation, stylization, video looping, color-block manipulation,

Excited to present VideoTree🌲 at #CVPR2025 Fri at 10:30AM! VideoTree improves long-video QA via smart sampling: -Query-adaptive: finds the parts of the video relevant to the query -Coarse-to-fine structure: structured hierarchically to sample granularly from relevant segments

🎉Excited to share that I’ll be starting my CS PhD journey at UNC-Chapel Hill UNC Computer Science this fall! 🎓 I’ll be working with the renowned Mohit Bansal at UNC NLP — a dream comes true! ✨ Huge thanks to everyone who's helped me get here. Can't wait to begin this new life and research journey! 🧳🚀