🔥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

Generative models (such as Dall-E 2 and PaLM) are becoming just such an insanely powerful, almost magic-like technology, it's completely NUTS. And it seems like most (non-ML) people still don't fully grasp the implications. This technology will thoroughly transform society.

Evaluating subgroup disparities in ML shouldn't only be a "fairness" thing. It should be ML/data science best practice in general. If your model performs poorly for some identifiable subgroup, you should probably go fix that, just in the interest of making a better model.

The preliminary list of accepted #NeurIPS2022 workshops is out! No matter whether you are attending in person or virtually, we offer a wide range of relevant workshops blog.neurips.cc/2022/07/17/ann…

Here's hoping I don't need to update this slide again before my talk next week EMNLP 2025 If anyone is planning to release anything next week, please lmk soon 😅 Am I missing any text-only LLMs?

It's deeply depressing that the killer app of contemporary machine learning is bullshit generation.

We need to stop conflating open/gated access and opensource. ChatGPT is *not* open source -- we don't know what model is under the hood, how it works, or any other tweaks/filters that are applied. (1/n)

ChatGPT is incredibly limited, but good enough at some things to create a misleading impression of greatness. it's a mistake to be relying on it for anything important right now. it’s a preview of progress; we have lots of work to do on robustness and truthfulness.

Excited to share our work on data minimization for ML! The principle of data minimization is a cornerstone of global data protection regulations, but how do we implement it in ML contexts? 🧵: Let's dive into some insights. 🔗: arxiv.org/abs/2405.19471

Are LLM adversaries more powerful than we assume? 🤔 Yes! Our work shows real-world attackers with multifaceted access to models and prompts have significantly higher extraction risks. An exciting paper led by Yash More, and with Golnoosh Farnadi @NeurIPS2024 arxiv.org/abs/2407.02596

What's the best bias mitigation algorithm? There isn't one! Benchmarks often miss critical nuances. In our latest paper, we unpack these pitfalls. Critical work w/ Usman Gohar , Lu Cheng , Golnoosh Farnadi @NeurIPS2024 🔥Spotlight AFME 2024 @ NeurIPS @ NeurIPS 2024 📰arxiv.org/pdf/2411.11101

1/ 🚨 Large Language Models (LLMs) like ChatGPT are changing our daily workflows, but what are the hidden costs? At COLM 2024, we presented our work on sensitive-disclosures in user-chatbot interactions. Here's a thread unpacking the key takeaways if you missed our poster! 🧵👇

🆘Help needed! Are you working on Privacy (from a Technical (e.g., Differential Privacy), Policy, or Law perspective)? Please give your availability to review for PPAI (ppai-workshop.github.io) if you can! We'd highly appreciate it! 🙏 forms.gle/dqjVsBsR2y81v1…

1/ 🚨 How vulnerable are large language models (LLMs) to extraction attacks? Our paper (presented at ACL PrivateNLP) explores a more realistic adversary capable of exploiting LLM brittleness and multi-faceted access to data. The risks might be bigger than you think. 🧵👇

TODAY it is time for the Algorithmic Fairness Workshop #AFME2024 at #NeurIPS2024! 📍West Meeting 111-112! Excited for our 5 contributed spotlight talks today! Benjamin Laufer Danny To Eun Kim (@teknology.bsky.social) Alex Tamkin Prakhar Ganesh Natalie Mackraz

😡 Absolutely disappointed with Overleaf. My account was deleted without my knowledge, and they’ve done nothing to help me recover it or transfer to my secondary email. Years of work, including all my CVs, SOPs, papers, etc., gone! This is unacceptable. #Overleaf