24 citations · 54 across the 17 of their papers we have counts for
20 papers
Retrieval-Centric Deep Learning in Growing Nonparametric Neural Networks
Maximilian Schlegel, Rajai Nasser, Seijin Kobayashi +6
We investigate a general-purpose layer for deep learning that, instead of compressing arbitrary-size training data into fixed-size weight matrices, stores a new pair of key-value r…
A game theory for foundation models shows new paths to rational cooperation through similarity inference
Alexander Meulemans, Maciej Wołczyk, Marissa A. Weis +11
As autonomous agents powered by foundation models are increasingly integrated into social and economic systems, understanding the principles governing their collective behavior is…
Emergent temporal abstractions in autoregressive models enable hierarchical reinforcement learning
Seijin Kobayashi, Yanick Schimpf, Maximilian Schlegel +12
Large-scale autoregressive models pretrained on next-token prediction and finetuned with reinforcement learning (RL) have achieved unprecedented success on many problem domains. Du…
Embedded Universal Predictive Intelligence: a coherent framework for multi-agent learning
Alexander Meulemans, Rajai Nasser, Maciej Wołczyk +10
The standard theory of model-free reinforcement learning assumes that the environment dynamics are stationary and that agents are decoupled from their environment, such that polici…
MesaNet: Sequence Modeling by Locally Optimal Test-Time Training
Johannes von Oswald, Nino Scherrer, Seijin Kobayashi +14
Sequence modeling is currently dominated by causal transformer architectures that use softmax self-attention. Although widely adopted, transformers require scaling memory and compu…
Weight decay induces low-rank attention layers
Seijin Kobayashi, Yassir Akram, Johannes Von Oswald
The effect of regularizers such as weight decay when training deep neural networks is not well understood. We study the influence of weight decay as well as -regularization whe…
Multi-agent cooperation through learning-aware policy gradients
Alexander Meulemans, Seijin Kobayashi, Johannes von Oswald +6
Self-interested individuals often fail to cooperate, posing a fundamental challenge for multi-agent learning. How can we achieve cooperation among self-interested, independent lear…
Learning Randomized Algorithms with Transformers
Johannes von Oswald, Seijin Kobayashi, Yassir Akram +1
Randomization is a powerful tool that endows algorithms with remarkable properties. For instance, randomized algorithms excel in adversarial settings, often surpassing the worst-ca…
When can transformers compositionally generalize in-context?
Seijin Kobayashi, Simon Schug, Yassir Akram +5
Many tasks can be composed from a few independent components. This gives rise to a combinatorial explosion of possible tasks, only some of which might be encountered during trainin…
Attention as a Hypernetwork
Simon Schug, Seijin Kobayashi, Yassir Akram +2
Transformers can under some circumstances generalize to novel problem instances whose constituent parts might have been encountered during training, but whose compositions have not…
Discovering modular solutions that generalize compositionally
Simon Schug, Seijin Kobayashi, Yassir Akram +6
Many complex tasks can be decomposed into simpler, independent parts. Discovering such underlying compositional structure has the potential to enable compositional generalization.…
Uncovering mesa-optimization algorithms in Transformers
Johannes von Oswald, Maximilian Schlegel, Alexander Meulemans +10
Some autoregressive models exhibit in-context learning capabilities: being able to learn as an input sequence is processed, without undergoing any parameter changes, and without be…
Gated recurrent neural networks discover attention
Nicolas Zucchet, Seijin Kobayashi, Yassir Akram +4
Recent architectural developments have enabled recurrent neural networks (RNNs) to reach and even surpass the performance of Transformers on certain sequence modeling tasks. These…
Would I have gotten that reward? Long-term credit assignment by counterfactual contribution analysis
Alexander Meulemans, Simon Schug, Seijin Kobayashi +2
To make reinforcement learning more sample efficient, we need better credit assignment methods that measure an action's influence on future rewards. Building upon Hindsight Credit…
Disentangling the Predictive Variance of Deep Ensembles through the Neural Tangent Kernel
Seijin Kobayashi, Pau Vilimelis Aceituno, Johannes von Oswald
Identifying unfamiliar inputs, also known as out-of-distribution (OOD) detection, is a crucial property of any decision making process. A simple and empirically validated technique…
Meta-Learning via Classifier(-free) Diffusion Guidance
Elvis Nava, Seijin Kobayashi, Yifei Yin +2
We introduce meta-learning algorithms that perform zero-shot weight-space adaptation of neural network models to unseen tasks. Our methods repurpose the popular generative image sy…
The least-control principle for local learning at equilibrium
Alexander Meulemans, Nicolas Zucchet, Seijin Kobayashi +2
Equilibrium systems are a powerful way to express neural computations. As special cases, they include models of great current interest in both neuroscience and machine learning, su…
Learning where to learn: Gradient sparsity in meta and continual learning
Johannes von Oswald, Dominic Zhao, Seijin Kobayashi +4
Finding neural network weights that generalize well from small datasets is difficult. A promising approach is to learn a weight initialization such that a small number of weight ch…
Posterior Meta-Replay for Continual Learning
Christian Henning, Maria R. Cervera, Francesco D'Angelo +6
Learning a sequence of tasks without access to i.i.d. observations is a widely studied form of continual learning (CL) that remains challenging. In principle, Bayesian learning dir…
Neural networks with late-phase weights
Johannes von Oswald, Seijin Kobayashi, Alexander Meulemans +3
The largely successful method of training neural networks is to learn their weights using some variant of stochastic gradient descent (SGD). Here, we show that the solutions found…