Pinned
Furong Huang
2,250 posts
Furong Huang
@furongh
Associate prof of computer science at University of Maryland. Researcher in #AI/#ML, AI #Alignment, #Trustworthy ML, #EthicalAI, AI for ALL.
In 2010, I came to the U.S. straight from undergrad for a PhD. Fifteen years later the map looks messy, but the line of best fit is clear. 🤍 2010 — PhD Year 1: my advisor said, “Take the ML course.” I had never heard of ML. With the most supportive, inspiring advisor, I pivoted
- I saw a slide circulating on social media last night while working on a deadline. I didn’t comment immediately because I wanted to understand the full context before speaking. After learning more, I feel compelled to address what I witnessed during an invited talk at NeurIPS 2024It is just so sad that the #NeurIPS2024 main conference ended with such a racist remark by a faculty when talking about ethics. How ironic! I also want to commend the Chinese student who spoke up right on spot. She was respectful, decent, and courageous. Her response was
00:00
- 🎬 Just like Nolan's 'Memento' rewrote storytelling, we're reshaping AI! Introducing '🎞️ Mementos': our benchmark pushing AI to understand sequences of images, not just stills. A real game-changer in AI's narrative. mementos-bench.github.io #AIStorytelling #Multimodal #LLMs #GenAI
- NeurIPS rebuttal deadline is around the corner 😬 I’m not an expert, but thought I’d drop my two cents on how to write a good rebuttal, especially for folks writing their first few. Hope this helps someone! 🧵👇 (And please chime in with your own tips; let’s crowdsource the
- Ever wondered if we can protect AI from attacks without sacrificing its everyday effectiveness? Our #ICLR2024 spotlight paper unveils how. Dive into the future of resilient, uncompromised AI with us! #RobustAI #PerformanceMatters Project page 🔗: protected-beyond-worst-case.github.io/home/ A🧵👇
00:00 - 🧠💡 What if your 7B model could beat GPT-4o and Qwen2.5-72B—using just 11k training samples? No distillation. No warm-start. Just smart data and reinforcement learning. Inspired by Moravec’s Paradox, we let the model decide what's actually hard. 🚨 New paper: "SoTA with Less:
00:00 - 🚀 LLMs need better reward models! 🚀 Current reward models are slow & inefficient—they score full responses AFTER generation, making test-time alignment painfully slow. What if we could guide LLMs token-by-token—in real time? 🔴 Meet GenARM! Our #ICLR2025 work introduces
- I don't usually share personal matters here, but I feel unsafe and need your help. This morning, I was left feeling deeply shaken. My neighbor, who routinely blocks my driveway, verbally abused and threatened me when I asked him to move his vehicle. This disruption has happened
- Deeply humbled to share my tenure promotion. This feels surreal and would not have been possible without the incredible support from so many people. To my mentors, colleagues, collaborators, and students - thank you for believing in me, inspiring me, and being there every step ofCongratulations to our eight faculty members who are receiving academic promotions in recognition of their continued excellence in research, scholarship and service to the @UofMaryland campus community and beyond. Learn more: go.umd.edu/1p8w
- 🚨 Breaking Research Discovery! 🚨 Large Vision Language Models (#VLMs) amaze with coherence but hide risks. 🤯🗡️ 🔍 Meet "Shadowcast"🥷: A stealthy, mind-bending AI data poisoning method. 🕵️♂️💻 Project page 🔗: vlm-poison.github.io #LLMs #DataSecurity A 🧵 👇
