Ilya Sutskever confirmed at NeurIPS '24 pre-training scaling law will not be the same. Internet data, the AI fossil fuel, is running out. It is not the end of the AI hype, the rapid rise of foundational AI has outpaced our ability to fully grasp their long-term societal impact. AI agents and applications are poised to transform every facet of our lives, and we will see it happening in coming few years. There will still be improvements in model architecture, and synthetic data creation, so models will improve, just not at the lightning speed they did till now. Looking forward, the single most important milestone in AI is creating systems capable of advancing themselves—AI researchers that can innovate and improve foundational models autonomously. We need to assess how far are we from this stage.

9

Ilya Sutskever at NeurIPS confirms: pre-training scaling for LLMs has stalled. Compute scales, but data doesn’t—neither new nor synthetic data is moving the needle. Alexandr Wang from Scale AI agrees: "The data wall is real, and synthetic data isn’t delivering." The next breakthroughs might come from focusing less on scaling and more on orchestration—harnessing multiple models and tools to solve complex, real-world problems collaboratively. What are your thoughts #ArtificialIntelligence #FutureOfAI #AITrends #LLMs

5

1 Comment

This is a fascinating paper evaluating DeepMind’s AlphaFold 3 and its approach to predicting ligand docking: https://lnkd.in/gieM2nW5 The authors used a clever method to evaluate AlphaFold, finding that it can make glaring mistakes when placing molecules in binding sites because its model is likely only fitted to the protein’s backbone geometry instead of actually learning the underlying physical interactions. Not that AlphaFold isn’t a breakthrough technology for drug discovery, but this paper highlights that even the most cutting-edge applications of AI are perhaps still best used to generate candidates, but humans still need to perform robust tests to validate them. We’ve been hearing a similar sentiment from nearly all of our users when bringing AI into cloud infrastructure management. Flightcrew has been read-only from the start, using AI to generate insights but requiring a human to sign off before any changes are applied to their infrastructure. The goal is to speed up existing workflows, not create a new point of failure. Drug discovery and managing Kubernetes are completely different applications but we’re all similarly wary of trusting full AI automation and at this rate, likely will be for some time.

19

One new interesting application of LLMs: CAIS has released a forecasting bot that provides probability estimates for any question you ask. For example, to the question "Will GPT5 be released before 2025?", it answers that it's 70% likely. The bot works by searching for relevant information online, feeding it to GPT-4o, asking a few guiding questions, and then requesting a probability estimate. What's impressive is that this bot outperforms experienced human forecasters, even though GPT-4o wasn't specifically trained for this task. This kind of application shows that AI could be of great help in improving decision-making in the future. Link to the blog post: https://lnkd.in/gjP7JQgz Link the bot if you want to try it yourself: https://forecast.safe.ai/

11

🎯Data was, is, and will be the only fuel. Analogy with the fossil fuel is a good one, since the data we have today is as dirty as oil and coal. What big tech companies tried is to put a lot of "garbage" data, and what they did get was fantastic. Yet, we are at the point of diminishing returns. Data is the beginning and the end of this cycle.” Great thread, particularly this comment 👆from Ilija Lazarevic

20

2 Comments

Today, I found a fascinating way to think about “Tokens” in LLMs. It came from a conversation between Jensen Huang and Patrick Collison at a recent Stripe event. When asked about the future of compute capacity in five years, Jensen smartly dodged that question a bit, but then gave an excellent explanation and analogy about why tokens are the new forces that will power the next decades of humanity. He explained that we're now producing something unprecedented (and at scale): floating point numbers that possess “value”, which we now call tokens. These tokens are valuable because they encapsulate “intelligence”. People are now taking these tokens and transforming them into English, French, images, videos, chemicals, proteins, robotic movements, etc. And many are working hard to expand the range of concepts and ideas we can create with these tokens. He then goes on to made a compelling comparison between tokens and electricity. In the past industrial revolution, we successfully found a way to convert “atoms” into “electrons” (by boiling water to power electricity turbines). And now, we've discovered a way to convert "electrons" into "tokens" (by using energy to power data centers that train and run LLMs). When electricity was first introduced, few people understood its value. Today, paying for kilowatts is routine. The same will happen with tokens. Right now, only early adopters and builders are paying tokens. Soon, everyone will be paying for tokens on a daily basis to supercharge productivity and power new products and services. Many new industries will be born from and built on top of tokens. When I first heard about this way of thinking, I got goosebumps. Perhaps it’s because Jensen has a good storytelling skill that helps him sell what he builds, but I have to admit this way of thinking gave me a brief surge of pride. I'm proud of humanity's collective effort and creativity, which have taken us from living under rocks to building machines that can "think." That's nothing short but a miracle.

18

5 Comments

Thanks Hamdi Amroun, PhD - this is a great paper to share. In the real world business context we often need AI to reason over large pieces of text, and often you need/want multi-document summarization (MDS). 'When testing five LLMs with benchmarks using news and conversation datasets, they found up to 75% of content in MDS summaries was hallucinated, with notable increases towards summaries' end. Alarmingly, even for non-existent topics, models like GPT-3.5-turbo and GPT-4 generated fabricated content 79% and 44% of the time, respectively. Analysis of 700+ generated insights showed that hallucinations often arose from failures to follow instructions or overly generic content.' I agree with Hamdi's sentiments that despite improvements in post-processing techniques, robust solutions are urgently needed to reduce hallucinations, as inaccurate summaries can lead to business risks (notably misinformation and misrepresentation). I would add that there is still a lot of utility in LLMs/Multimodal models but we need to be vigilant - this is not about AI malevolence or intentional deception. LLMs do not possess consciousness or intent; they generate content based on patterns in data they were trained on.

5

1 Comment

Yes, the future is AI Agents navigating the Web in the decentralised manner, using Croissant for Machine Learning to check content ownership and take licenses into account. It's a kind of "TomTom" for the future Web, represented as moving knowledge graph. We knew all of that for years and it's coming.

2

1 Comment

💡 ByteDance's Latest Work on Neural Network-Based Molecular Force Fields 💡 I recently read ByteDance’s work on developing a new force field, and it’s pretty good! 👍 ByteFF is a molecular mechanics force field trained on a massive dataset: 2.4 million geometries and 3.2 million torsion profiles. It leverages GNN to predict bonded and non-bonded parameters, showing strong performance in predicting molecular geometry, torsion energy, and conformational energy. This model has great potential for drug discovery and molecular simulation. I do have some concerns about using torsion scan + hessian methods to train non-bonded parameters in this framework. Non-bonded interactions often involve long-range effects, which are challenging for GNNs to capture accurately. That said, the overall results presented in this work are still quite promising. It's exciting to see tech giants like ByteDance stepping into AI for Science, and their vast resources enable them to support DFT calculations on this massive scale. (🤞 Hopefully, they will release the dataset for open use, LOL!) 👉 You can find the paper here: https://lnkd.in/gjmGV4fd 👉 And check out their open-source framework on GitHub, which is well-engineered: https://lnkd.in/gkxcTi6t #AIforScience #DrugDiscovery #MolecularModeling #ForceField #ByteDance

49

1 Comment

Great to see the newly announced Bamba-9B, an inference-efficient Hybrid Mamba2 model 🐍 trained by IBM, Princeton, CMU, and UIUC on completely open data used PyTorch FSDP to train these novel architecture models and they are integrating inference with vLLM, which recently joined the PyTorch Ecosystem.

167

Implementing Reinforcement Learning (RL) in the context of Large Language Models (LLMs) posed challenges, particularly in designing reward engineering and penalization components. In traditional RL, reward engineering is straightforward, often defined through a scalar reward function. However, implementing the same concept in RL with LLMs required a critic model that evaluates the questions and answers, critiquing them in a manner that simulates the reward function. A recent research paper on Natural Language Reinforcement Learning (NLRL) provides an intriguing perspective on the reward problem in RL by translating traditional RL concepts into natural language constructs. Here's a summary of how it addresses the reward problem: Language-Based Value Functions: Unlike traditional RL, which uses scalar rewards, NLRL employs language-based value functions. These functions articulate the effectiveness of policies and actions in a more interpretable, natural language format. This approach addresses the need for richer, more informative feedback than simple numerical scores. Iterative Refinement of Language Policies and Critics: NLRL incorporates a structured training pipeline where language-based policies and critics are iteratively refined using textual feedback. This method allows the system to handle the complexities of defining rewards in natural language, improving its understanding and evaluation of actions over time through textual feedback loops. Use of Large Language Models (LLMs): The framework leverages the advanced capabilities of LLMs to interpret, generate, and evaluate complex language-based information. This is essential for implementing language-based value functions and policies, enabling the system to generate and refine responses that align closely with desired outcomes. Distilling Enhanced Actions and Evaluations: NLRL employs a process where enhanced evaluations from the language-based value function approximator and improved actions from the language policy improvement operator are distilled. This allows for continuous refinement and learning. Compatibility with Existing LLMs: The framework is designed to be compatible with current LLMs, enabling it to leverage pre-existing language models to process and generate linguistic data effectively. This compatibility is crucial for translating the traditional reward engineering process into a language-based framework. The NLRL framework's introduction of natural language as a medium for representing and evaluating policies and rewards offers a novel solution to the reward problem in RL. By shifting from numerical reward systems to a more flexible, expressive, and intuitive system based on language, NLRL enhances interpretability and adaptability, paving the way for more diverse RL applications. Innovation Hacks AI Inc. #llm #refincorcementlearning #ai

4

Amidst a range of commercial black box LLMs from large enterprises, we are also seeing Open-Source Machine Learning Frameworks for Advancing Large Language Models in Healthcare. Check out Hippocrates, an open-source LLM framework specifically developed for the medical domain https://lnkd.in/gw5xHdUD

1

Ilya like all other LLM vendors is now actively trying to cover up the 10s of billion$ in LLM tech debt by claiming “Agentic AI” as the new panacea to solve intelligence - even lowering the bar for AGI. You can't brute force your way to solve intelligence. There is no "i" in AI without "Cognition". Self-learning and adapting in real-time isn't going to happen with LLMs or the wrapper "Agents" sitting on top of these LLMs. Hopefully they are not going to sit on "Agentic AI" for another few years to arrive at this conclusion as they did with the scaling laws. We need a Neuro-Symbolic approach centered around human-like Cognition to solve intelligence that doesn’t require trillions of data points, billion$ in computational resources and even more $$ in energy requirements while it cannot learn & adapt continuously yet after all this it hallucinates. INSA is the exact opposite of this. INSA: Integrated Neuro-Symbolic Architecture https://lnkd.in/ggFTxuUz The insanity of huge language models: https://lnkd.in/d5pcyf6R

32

11 Comments

We recently posted a new preprint "Entanglement-enhanced learning of quantum processes at scale" (https://lnkd.in/eCeFCrF5) with Senrui Chen, Swarnadeep Majumder, Haoran Liao, Derek S. Wang, Moein Malekakhlagh, Ali Javadi, Liang Jiang, and Zlatko Minev. Check out the nice summary by Senrui below (https://lnkd.in/eWBkCAGS): "Can (noisy) quantum entanglement provide advantages in learning physical processes of practical interest? We give a positive answer to this question, in the task of learning the Pauli noise processes on a quantum device. Pauli noise learning is a well-studied task in the literature of quantum noise characterization. Recent works have proven rigorous efficiency enhancement in Pauli channel learning using entanglement with quantum memory, but it remains unclear whether such enhancement can survive noise and be practically useful. Inspired by techniques from quantum benchmarking and error mitigation, we develop an entanglement-enhanced Pauli channel learning scheme that is both noise-resilient and provably efficient. Experimental results with IBM Quantum confirm our protocol yield consistent estimates with standard entanglement-free learning schemes, while possesses a significant improvement in sample efficiency, even on current noisy quantum hardware. Our work showcases quantum entanglement as a resource are already giving us enhancement on current quantum devices. Our protocol also opens up new possibility for quantum noise characterization at scale."

105

3 Comments

Today's guest is the amazing Google DeepMind research scientist, Dr. Rosanne Liu (pictured here in San Francisco this past weekend with myself and Jason Yosinski, who introduced me to Rosanne). Rosanne:  • Is a Research Scientist at Google DeepMind in California. • Is Co-Founder and Executive Director of ML Collective, a non-profit that provides global ML research training and mentorship. • Was a founding member of Uber AI Labs, where she served as a Senior Research Scientist. • She has published deep learning research in top academic venues such as NeurIPS, ICLR, ICML and Science, and her work has been covered in publications like WIRED and the MIT Tech Review. • Holds a PhD in Computer Science from Northwestern University. Today’s episode, particularly in the second half when we dig into Rosanne’s fascinating research, is relatively technical so will probably appeal most to hands-on practitioners like data scientists and ML engineers. In today’s episode, Rosanne details:  • The problem she founded the ML Collective to solve. • How her work on the “intrinsic dimension” of deep learning models inspired the now-standard LoRA approach to fine-tuning LLMs. • The thorny problems with LLM evaluation benchmarks and how they might be solved. • The pros and cons of curiosity- vs goal-driven ML research. • The positive impacts of diversity, equity and inclusion in the ML community. The "Super Data Science Podcast with Jon Krohn" is freely available on all major podcasting platforms and we publish a video version on YouTube. This is Episode #797! #superdatascience #machinelearning #ai #deeplearning #llms

114

11 Comments

🚀 Excited to share that our latest work, S-MolSearch, has been accepted at NeurIPS 2024! 🚀 The paper, "S-MolSearch: 3D Semi-supervised Contrastive Learning for Bioactive Molecule Search", presents a novel semi-supervised learning framework designed to tackle the challenge of bioactive molecule search by effectively combining labeled and unlabeled data. By using soft labels generated through inverse optimal transport, this method enables the model to learn from a vast amount of unlabeled data, which is particularly valuable in scenarios where obtaining labeled data is difficult or expensive—such as in early drug design. 🌟 I find that S-MolSearch's approach has a lot in common with the self-distillation method applied in AlphaFold3. Just as AlphaFold3 improved its performance by learning from AlphaFold2 predictions, S-MolSearch's soft labels guide the model to better utilize both labeled and unlabeled molecular data. I believe this semi-supervised approach has huge potential for further exploration in the Molecular Representation space. 🧠 👉 Check out our paper here: https://lnkd.in/gJiGg2DE 🎉 Congratulations to Zhifeng Gao for this work! 🎉 It’s been exciting to follow the progress of this work and discuss on it. #AIforScience #UniMol #AlphaFold #AlphaFold3 #DrugDesign #MolecularRepresentation #NeurIPS2024

44

2 Comments

Recently I learned that LLMs show ability to acquire rare linguistic competence (e.g. "a beautiful five days" is acceptable, but "a blue five pencils" not so much), from a tech talk by Prof. Kyle Mahowald from UT Austin (thanks Dr. Steve Kramer for organizing the guest speaker series). A small, high quality dataset BabyLM was sufficient to study the effect of training data on the model's linguistic ability. By surgically removing different types of phases from training data and observing the degradation model's ability, it can be revealed which part of data is responsible for the particular linguistic ability. This is a great example of doing AI research without breaking a bank. As someone not familiar with linguistics, I take linguistic competence to be a linguistic "value function" or "reward function". Can similar emergence of rare competence (not in training data but model learned it anyway) be shown in other domains? Reference: https://lnkd.in/g4cAPUYk

16

1 Comment

in the age of big data... more data are needed!

11