Representation learning

…or JAX, probability, statistics and representation learning are required; neuroimaging and large-scale computing experience are desirable. Appointment: 3 years; advertised start 2027-01-01

…including continual and adaptive learning, forgetting mitigation, representation learning, explainability, neuromorphic computation, computational neuroscience and multimodal models. Research papers connect theoretical methods with applications across science and engineering. Review is double-blind. Standard submissions use six IEEE-format pages; up to four extra pages incur fees. Accepted authors must submit final papers and pay the corresponding registration fees by 12 April 2027, 23:59 AoE. Remote authors may supply prer

…includes mechanistic interpretability, representation learning and alignment, biological and artificial intelligence, neural dynamics, and models of perception, cognition and action. The primarily single-track format combines keynotes, research talks and posters with some parallel community sessions. Generative Adversarial Collaborations examine competing theories; keynote-and-tutorial sessions connect

…Research covers signal processing, representation learning, motor-intent decoding, object and hand understanding, multimodal reasoning and real-time sensing. The role focuses on robotics and interactive systems; it has no clinical or biomedical focus. Candidates may also prototype wearable hardware and evaluate mixed-reality applications.

Representation learning underlies visual search, retrieval and recognition by encoding gallery images and matching them with queries. Most early work assumes a fixed model, but new training data or more expressive architectures can change the feature representation. Recomputing every gallery vector after an update is expensive for collections containing billions of images and may be impossible when the original images are unavailable because of privacy or storage restrictions. Compatible representation learning aims to update the model while retaining usable existing gallery features. The talk

…Recent approaches in representation learning circumvent the need for labeled examples, and match or surpass supervised learning methods on a variety of tasks. These approaches generally rely on supervision signals extracted from the

Identifying and characterizing the wiring rules of neural circuits is critical for our understanding of the computations that they perform and how they learn. The reconstruction of complete connectivity maps of neural networks at the single cell resolution make the quantitative study of their design principles possible. We examined the detailed connectivity of a column of the mouse visual cortex with a generative model that predicts connectivity based on explicit biological and physical features: excitatory or inhibitory cell type, the distance between cell bodies, and cell's depth, which is c

…artificial neural networks that learn representations by incorporating brain-like learning abilities. Recent work has focused on learning representations (latents) by capturing similarity between inputs and latents, but these methods have typically been used to learn downstream features from the input and have not been studied in the context of a generative paradigm, where one can map the representations back to the input space, incorporating not only bottom-

Integrating task-relevant information into neural representations is essential for both biological and artificial intelligence. However, current experimental limitations in tracking synaptic weight changes make it challenging to understand learning mechanisms in animals. To address this, we propose a framework based on neural manifold geometry to study representation learning directly from neural activity. A neural manifold represents the collection of population responses to the same task condition (e.g., a cat manifold for responses to images of cats) and can be estimated from recorded activ

…subjects generalize in part through learning a representation that encodes the compositional structure (Whittington et al., 2020; It

Developing scientifically-grounded theories for representation learning and generalization in artificial deep neural networks remains a grand challenge of fundamental interest to theoretical neuroscience and machine learning. I will discuss our work on one facet of this challenge — namely understanding generalization or “scaling laws” in learned neural networks as a function of basic control variables. I’ll discuss a taxonomy we develop that classifies different regimes of scaling behavior. We identify regimes where generalization exhibits universal scaling behavior and others where it can be

…The learning rule and initial synaptic weights determine how well a network will generalize on new data. Here, we investigate how initial synaptic variability and representation learning affect the learning curves of neural networks (generalization error vs sample size). We show that two scaling regimes hold for many plausible learning rules including feedback alignment (FA), direct feedback alignment (DFA) and Hebb rules, as well as implausible gradient descent (GD). In the sm

…Classical models of representation learning assume that feedforward weights are learned via pairwise Hebbian-like plasticity. Here, we show that pairwise Hebbian-like plasticity only works under unrealistic requirements on neural dynamics and input statistics. To overcome these limitations, we derive from first principles a learning scheme based on voltage-dependent synaptic plasticity rules. Here, recurrent connections learn to locally balance feedforward input in indivi

Children learn to build a visual representation of the world from unsupervised exploration and we hypothesize that a key part of this learning ability is the use of self-generated navigational information as a similarity label to drive a learning objective for self-supervised learning. The goal of this work is to exploit navigational information in a visual environment to provide performance in training that exceeds the state-of-the-art self-supervised training. Here, we show that using spatial and temporal information in the pretraining stage of contrastive learning can improve the performan

…sustain stable memories while flexibly learning new associations, allowing animals to adapt to dynamic environments while executing precise behaviors years after they were first learned. Yet, recent experiments have revealed that neural representations of fixed stimuli change continuously over time, contravening the classical assumption that learned features should remain static to maintain task proficiency. The phenomenon of representational drift can be reconciled with normative principles for neural computation within the framework of pro

…Classical models of representation learning assume that input weights are learned via pairwise Hebbian-like plasticity. Here, we show that pairwise Hebbian-like plasticity only works under specific requirements on neural dynamics and input statistics. To overcome these limitations, we derive from first principles a learning scheme based on voltage-dependent synaptic plasticity rules. Here, inhibition learns to locally balance excitatory input in individual dendritic compa

…intelligence through improved representation learning. Before joining DeepMind, Irina was a British Psychological Society Undergraduate Award winner for her achievements as an undergraduate student in Experimental Psychology at Westminster University, followed by a DPhil at the Oxford Centre for Computational Neuroscience and Artificial Intelligence, where she focused on understanding the computationa