Severin Hacker

Startups are very similar to research labs Over the last few months Ankit Sanghvi, Jayanth Krishnaprakash, and I have found ourselves using experimental terms more than ever: Drafting hypotheses Designing experiments Running A/B tests Controlling variables We're focused on testing our main idea: "Can AI effectively collect debt?" The toughest part is staying aware of and reducing our biases every step of the way.

64

10 Comments

Can LLMs strengthen LP relaxations and speed up solver convergence on MILPs? This is a question that we (Milad Yazdani, Samin Aref, Zirui Zhou, and I) explored in our recent work, and the results are promising. In this year’s IMO, leading AI systems (from OpenAI, Google/DeepMind) achieved a gold-medal–level score by solving 5 out of 6 problems, yet P6 (a notoriously difficult problem) remained unsolved. We modeled P6 as an integer program (LLMs, with some guidance, also produced correct formulations). Still, top MILP solvers did not converge in a reasonable time, suggesting a weak LP relaxation. Our method: EvoCut. EvoCut combines evolutionary search + LLMs to automatically discover problem-specific cuts (valid inequalities) that strengthen the relaxation. Results: - ~55% average gap reduction at the same runtime - Order-of-magnitude faster convergence in many instances We also observed similar improvements on other combinatorial problems across different domains. Read more & try EvoCut: - Paper: EvoCut: Strengthening Integer Programs via Evolution-Guided Language Models (https://lnkd.in/g6Kydtp6) - Code: Test EvoCut on your own MILP (https://lnkd.in/g42wMy4s) #MILP #OperationsResearch #Optimization #AI #LLMs #CuttingPlanes #Combinatorics #IMO #LinearProgramming

123

4 Comments

Unpopular opinion: The RL hype is a red flag for technical depth. When I hear founders or VCs casually dropping "reinforcement learning" in pitches, it's usually code for "we don't actually understand our problem space." Here's the reality check: RL only works when you have: - Perfect simulation environments - Extremely fast iteration cycles - Clear reward functions - Tolerance for massive compute costs That's why it succeeds in games (perfect simulation) and search optimization (Massive compute capacity at Deepmind AlphaGo/Anthropic/OpenAI) but fails spectacularly in the messy, slow-feedback real world. The graveyard of "RL-powered" startups tells the story. They burned through $$$$ trying to apply algorithms to problems that needed simple heuristics or supervised learning. Most of the time, a well-tuned evolutionary algorithm will outperform your fancy RL setup while using <1/10th the compute and delivering results you can actually explain to stakeholders. The next time someone pitches you an RL solution, ask them: "Have you tried solving this with a basic optimization approach first?" Watch how quickly the conversation changes. I am surprised when people are surprised that Direct Preference Optimization (DPO) can exceed PPO-based RLHF [1] or Reflective Prompt Evolution Can Outperform Reinforcement Learning [2]. [1] https://lnkd.in/gCY3bcTQ [2] https://lnkd.in/gQRwyQyM

45

2 Comments

Axiom, a pre-product AI startup founded by a Stanford math PhD Carina Hong, is raising at a $300m valuation. Goal is to build AI trained on math proofs that can solve complex math problems with certainty, for Wall St. quants. Carina graduated from MIT with a double major in mathematics and physics. She then earned a neuroscience master’s degree with dissertation distinctions from Oxford, now pursuing a law degree and PhD in mathematics at Stanford.

63

3 Comments

🚀 Choosing the Right LLM Made Easy! A few days ago, DeepSeek AI made headlines for achieving top scores across multiple LLM benchmarks—competing with OpenAI, Google, and Anthropic. But here’s the thing… most of us don’t even know what these benchmarks really measure. 💡 Let’s break it down. What do LLM benchmarks actually test? 🔹 GLUE & SuperGLUE – How well an LLM understands and processes language. 🔹 MMLU & OpenBookQA – General knowledge and subject expertise. 🔹 GSM8K & AGIEval – Problem-solving and math skills. 🔹 CodeXGLUE & HumanEval – How well an LLM can write and test code. With so many AI models available, these benchmarks make it easier to choose the right one for your needs. 📌 Save & Share this post to help others in AI! ➕ Follow GetGenerative.ai Prateek Kataria for more AI insights 🚀 #AI #LLM #MachineLearning #ArtificialIntelligence #Tech #DeepLearning #AIResearch #DataScience #AITrends #NeuralNetworks #GenerativeAI #Automation #Innovation #TechTrends #AIForEveryone

43

How do LLMs really compare to the human brain? In a sequence of studies, our team at Google Research and partners at Princeton, HUJI and NYU have been exploring this question, aligning brain activity with the internal contextual embeddings of speech and language within LLMs as they process everyday conversations. What have we discovered? AI language models and the human brain share underlying organizational principles for representing words. Specifically, the internal representations used by these AI systems correspond to the neural activity patterns observed in brain regions crucial for both understanding and generating speech, hinting at potential shared computational mechanisms, all in all demonstrating the power of deep language model’s embeddings to act as a framework for understanding how the human brain processes language. Thank you to our research partners: Hasson Lab at Princeton University, the DeepCognitionLab at the Hebrew University, and researchers from the NYU Langone Comprehensive Epilepsy Center. Google Research Blog Post, authored by Mariano Schain and Ariel Goldstein: https://lnkd.in/d9cbP4Ee Nature Human Behaviour Paper: https://lnkd.in/dCbvdC-A Nature Communications Paper: https://lnkd.in/dbPCGY_m Nature Neuroscience Paper: https://lnkd.in/dvWjEg3C Hasson Lab, Princeton: https://lnkd.in/diMh8RA4 DeepCognitionLab at the Hebrew University: https://lnkd.in/dFPDuM7x  NYU Langone Comprehensive Epilepsy Center: https://lnkd.in/dU8rWe5e

680

26 Comments