#KDD2024 kicked off yesterday in Barcelona, and we are already on a fast start! Several School of CSE faculty, students, and alumni organized and presented today at the epiDAMIK Workshop workshop! Check out the website 🔗👇 for more on the workshop! epidamik.github.io/index.html

More exciting news from #KDD2024! A CSE/NASA JPL collaborative paper won the conference best paper runner-up! Congratulations ML Ph.D. student Austin Wright, Professor Polo Chau, and Scott Davidoff! Check out the paper on Nested Fusion here: dl.acm.org/doi/10.1145/36… Duen Horng "Polo" Chau

😎 Our paper on the LLM safety landscape has been accepted at NeurIPS Conference 2024! #Safety #LLM #MachineLearning

Please join us in congratulating longtime staff member, Queenie Kravitz, on her retirement today. She started Carnegie Mellon University in 1993 and the HCII in 2004, and as graduate program coordinator certified our very first HCI PhD and master's degrees. Congrats, Queenie! #CMUhcii

CSE Prof. Duen Horng "Polo" Chau and his group are presenting two papers and two posters this week at IEEE VIS! Check out the interactive graphic 🔗👇 for a peek of all Georgia Tech research presented this week, including award-winning work on Transformer Explainer! public.tableau.com/views/VIS2024/…

🚀Excited to present Diffusion Explainer at the IEEE VIS tomorrow at 1:45pm EST in the AI & LLM session! Try it now: poloclub.github.io/diffusion-expl… #StableDiffusion #GenerativeAI #AI #Visualization #IEEEVIS2024

Transformers visually explained: poloclub.github.io/transformer-ex…

🧑‍💻 The code of our NeurIPS'24 LLM safety landscape paper is now publicly available at: github.com/poloclub/llm-l… x.com/RealAnthonyPen…

🎉The coolest #CSE school in the world is hiring multiple faculty members! Application link below👇

🚀 Effective Guidance for Model Attention with Simple Yes-no Annotations Excited to share that I'll be presenting our recent work 🎨CRAYON🖍️ at IEEE Big Data soon! Catch me at 2pm in the Deep Learning II session!

Gradient descent alone tends to converge to local minima. Momentum frames optimization as a ball with mass moving down a hill. By adding inertia, the ball resists settling in small basins, allowing it to arrive at the global minimum.

Introducing ConceptAttention, an approach to interpreting diffusion transformer models! Write a prompt, choose some concepts, generate an image, and get high-quality heatmaps of text concepts. Our method outperforms existing methods like cross attention. Link to demo 👇

Diffusion Transformers aren't just generative models, but also powerful multi-modal encoders. ConceptAttention creates rich heatmaps of text concepts in images from DiT representations. This even works on real images, and can be applied to tasks like segmentation! Demo 👇

Create heatmaps that localize text concepts in generated videos. We discovered that our approach, ConceptAttention, can be directly extended from image generation to video generation models! It's amazing how simple techniques often generalize way better than more complex ones.

One of the simplest algorithms for sampling from a probability distribution is Random Walk Metropolis-Hastings. It proposes new samples by taking Gaussian-distributed steps, accepting or rejecting them to maintain the target distribution. I call this pdf the "fidget spinner".

This website has visualizations to understand almost all major topics in Machine Learning (link in comment)

Guardrail models like 🛡️ Llama Guard do more than filtering — we repurpose them to track how safety risk evolves 📉 through a response. This gives rise to the STAR ⭐ score: a fine-grained signal for finetuning LLMs more safely 🤖🔒 Curious how it works? More in the thread 👇

🚨 New work: We rethink how we finetune safer LLMs — not by filtering after the generation, but by tracking safety risk token by token during training. We repurpose guardrail models like 🛡️ Llama Guard and Granite Guardian to score evolving risk across each response 📉 — giving

🎉Our paper "Interpretation Meets Safety: A Survey on Interpretation Methods and Tools for Improving LLM Safety" has been accepted to EMNLP 2025 Main Track! EMNLP 2025 👉First survey connecting LLM interpretation & safety

Qwen Congrats on the great work! The "token-level safety detection" idea echoes our recent NeurIPS'25 dynamic safety shaping paper! 👉 arxiv.org/abs/2505.17196