Portrait de Razvan Pascanu

Razvan Pascanu

Membre affilié
Chercheur scientifique principal, Google DeepMind
Sujets de recherche
Apprentissage à quelques exemples
Apprentissage continu
Apprentissage de représentations
Apprentissage par renforcement
Apprentissage profond
Apprentissage profond géométrique
Apprentissage tout au long de la vie
Généralisation
Interprétabilité mécanistique
Optimisation
Réseaux de neurones
Réseaux de neurones en graphes
Réseaux de neurones profonds
Réseaux de neurones récurrents
Théorie de l'apprentissage automatique

Étudiants actuels

Maîtrise recherche - UdeM
Superviseur⋅e principal⋅e :

Publications

RAT: Bridging RNN Efficiency and Attention Accuracy in Language Modeling
Xiuying Wei
Anunay Yadav
Transformers have become the cornerstone of modern large-scale language models; however, their dependence on softmax attention poses a major… (voir plus) computational bottleneck, particularly in long-context settings. In this work, rather than following prevalent approaches such as linear attention (or SSMs) and local attention, we introduce an intermediate design called \rat between recurrence and attention mechanisms. It partitions the input into chunks, applies a simple linear recurrence within each chunk to capture local dependencies, and then performs softmax attention across chunks to model long-range interactions. By adjusting the size of the chunk, \rat enables flexible trade-offs, combining the strengths of RNN and attention. Empirically, with a chunk size of 16, the \rat layer achieves a \(7\times\) improvement in training speed with 100K token sequences and \(9\times\) in generation at 4K sequence length, while maintaining similar or sometimes even better accuracy compared to standard attention. We demonstrate this by training 1.3B parameter models from scratch and performing large-scale evaluations, including short- and long-context benchmarks, as well as supervised fine-tuning~(SFT). We further propose a hybrid architecture that interleaves \rat with local attention. By combining efficient long-range modeling with strong local interactions, this hybrid design not only improves inference speed and reduces cache memory usage compared to attention, but also consistently enhances performance, for example, achieving an average 1 point gain in commonsense reasoning tasks, up to 4 points on code tasks, and a 1 point Rouge-L increase in a summarization SFT task. Code is available at https://github.com/CLAIRE-Labo/RAT
How Overconfidence in Initial Choices and Underconfidence Under Criticism Modulate Change of Mind in Large Language Models
Dharshan Kumaran
Stephen M Fleming
Larisa Markeeva
Joseph Heyward
Andrea Banino
Mrinal Mathur
Simon Kayode Osindero
Benedetto De Martino
Viorica Patraucean
Large language models (LLMs) exhibit strikingly conflicting behaviors: they can appear steadfastly overconfident in their initial answers wh… (voir plus)ilst at the same time being prone to excessive doubt when challenged. To investigate this apparent paradox, we developed a novel experimental paradigm, exploiting the unique ability to obtain confidence estimates from LLMs without creating memory of their initial judgments -- something impossible in human participants. We show that LLMs -- Gemma 3, GPT4o and o1-preview -- exhibit a pronounced choice-supportive bias that reinforces and boosts their estimate of confidence in their answer, resulting in a marked resistance to change their mind. We further demonstrate that LLMs markedly overweight inconsistent compared to consistent advice, in a fashion that deviates qualitatively from normative Bayesian updating. Finally, we demonstrate that these two mechanisms -- a drive to maintain consistency with prior commitments and hypersensitivity to contradictory feedback -- parsimoniously capture LLM behavior in a different domain. Together, these findings furnish a mechanistic account of LLM confidence that explains both their stubbornness and excessive sensitivity to criticism.
Filter Equivariant Functions: A symmetric account of length-general extrapolation on lists
Owen Lewis
Neil Ghani
Andrew Joseph Dudzik
Christos Perivolaropoulos
MesaNet: Sequence Modeling by Locally Optimal Test-Time Training
Johannes Von Oswald
Seijin Kobayashi
Luca Versari
Songlin Yang
Maximilian Schlegel
Kaitlin Maile
Yanick Schimpf
Oliver Sieberling
Alexander Meulemans
Rif A. Saurous
Charlotte Frenkel
Blaise Agüera y Arcas
João Sacramento
Sequence modeling is currently dominated by causal transformer architectures that use softmax self-attention. Although widely adopted, trans… (voir plus)formers require scaling memory and compute linearly during inference. A recent stream of work linearized the softmax operation, resulting in powerful recurrent neural network (RNN) models with constant memory and compute costs such as DeltaNet, Mamba or xLSTM. These models can be unified by noting that their recurrent layer dynamics can all be derived from an in-context regression objective, approximately optimized through an online learning rule. Here, we join this line of work and introduce a numerically stable, chunkwise parallelizable version of the recently proposed Mesa layer (von Oswald et al., 2024), and study it in language modeling at the billion-parameter scale. This layer again stems from an in-context loss, but which is now minimized to optimality at every time point using a fast conjugate gradient solver. Through an extensive suite of experiments, we show that optimal test-time training enables reaching lower language modeling perplexity and higher downstream benchmark performance than previous RNNs, especially on tasks requiring long context understanding. This performance gain comes at the cost of additional flops spent during inference time. Our results are therefore intriguingly related to recent trends of increasing test-time compute to improve performance -- here by spending compute to solve sequential optimization problems within the neural network itself.
Unpacking Softmax: How Temperature Drives Representation Collapse, Compression, and Generalization
Wojciech Masarczyk
Mateusz Ostaszewski
Tin Sum Cheng
Tomasz Trzci'nski
Aurélien Lucchi
The softmax function is a fundamental building block of deep neural networks, commonly used to define output distributions in classification… (voir plus) tasks or attention weights in transformer architectures. Despite its widespread use and proven effectiveness, its influence on learning dynamics and learned representations remains poorly understood, limiting our ability to optimize model behavior. In this paper, we study the pivotal role of the softmax function in shaping the model's representation. We introduce the concept of rank deficit bias - a phenomenon in which softmax-based deep networks find solutions of rank much lower than the number of classes. This bias depends on the softmax function's logits norm, which is implicitly influenced by hyperparameters or directly modified by softmax temperature. Furthermore, we demonstrate how to exploit the softmax dynamics to learn compressed representations or to enhance their performance on out-of-distribution data. We validate our findings across diverse architectures and real-world datasets, highlighting the broad applicability of temperature tuning in improving model performance. Our work provides new insights into the mechanisms of softmax, enabling better control over representation learning in deep neural networks.
On the generalization of language models from in-context learning and finetuning: a controlled study
Andrew Lampinen
Arslan Chaudhry
Stephanie C.Y. Chan
Cody Wild
Diane Wan
Alexander Y. Ku
Alex Ku
Murray P. Shanahan
James L McClelland
LLMs are Greedy Agents: Effects of RL Fine-tuning on Decision-Making Abilities
Thomas Schmied
Jordi Grau-Moya
Why do LLMs attend to the first token?
Federico Barbero
'Alvaro Arroyo
Xiangming Gu
Christos Perivolaropoulos
Michael M. Bronstein
NoProp: Training Neural Networks without Back-propagation or Forward-propagation
Qinyu Li
Yee Whye Teh
How do language models learn facts? Dynamics, curricula and hallucinations
Nicolas Zucchet
Stephanie Chan
Andrew Lampinen
Soham De
Disentangling the Causes of Plasticity Loss in Neural Networks
Clare Lyle
Zeyu Zheng
Hado van Hasselt
James Martens
Will Dabney
From Markov to Laplace: How Mamba In-Context Learns Markov Chains
Marco Bondaschi
Nived Rajaraman
Xiuying Wei
Kannan Ramchandran
Michael C. Gastpar
Ashok Vardhan Makkuva