Guillaume Hennequin

Reciprocal connections dynamically build consensus between neocortical areas

Mitra Javadzadeh, Marine Schimel, Sonja B. Hofer, Yashar Ahmadian et autres

The neocortex is organized into specialized areas. Although computations within individual areas have been well studied, it is unclear how these regions function collectively and reconcile potential conflicts to form coherent percepts and decisions. We investigated the joint dynamics of primary (V1) …

us, gb, Soudan du Sud (code pays fourni par la source)

Improved denoising diffusion probabilistic models with efficient non-diagonal covariance modeling

Rui Xia, Ayan Das, Artem Artemev, Andi Zhang et autres

The sampling process of Denoising Diffusion Probabilistic Models (DDPMs) can be accelerated by leveraging second-order information in the form of approximations to the denoising posterior covariance -- allowing samples of acceptable quality to be produced in fewer but larger sampling steps. Previous …

Exploiting weight-space symmetries for approximating curvature

Artem Artemev, Rui Xia, Benjamin M. Boyd, Youjing Yu et autres

Many machine learning techniques rely on approximating a loss function's curvature, but this is notoriously hard to do at the scale of modern deep networks. Surprisingly, no previous work has exploited the curvature constraints that arise from well known weight-space symmetries in …

Exploiting weight-space symmetries for approximating curvature

Artem Artemev, Rui Xia, Benjamin M. Boyd, Youjing Yu et autres

Many machine learning techniques rely on approximating a loss function's curvature, but this is notoriously hard to do at the scale of modern deep networks. Surprisingly, no previous work has exploited the curvature constraints that arise from well known weight-space symmetries in …

Efficient Model Compression Techniques with FishLeg

Jamie McGowan, Wei Sheng Lai, Weibin Chen, Henry Aldridge et autres

In many domains, the most successful AI models tend to be the largest, indeed often too large to be handled by AI players with limited computational resources. To mitigate this, a number of compression methods have been developed, including methods that prune …

Dynamic consensus-building between neocortical areas via long-range connections

Mitra Javadzadeh, Marine Schimel, Sonja B. Hofer, Yashar Ahmadian et autres

Abstract The neocortex is organized into functionally specialized areas. While the functions and underlying neural circuitry of individual neocortical areas are well studied, it is unclear how these regions operate collectively to form percepts and implement cognitive processes. In particular, it remains …

us, gb (code pays fourni par la source)

Exact, Tractable Gauss-Newton Optimization in Deep Reversible Architectures Reveal Poor Generalization

Davide Buffelli, Jamie McGowan, Wangkun Xu, Alexandru Cioba et autres

Second-order optimization has been shown to accelerate the training of deep neural networks in many applications, often yielding faster progress per iteration on the training loss compared to first-order optimizers. However, the generalization properties of second-order methods are still being debated. Theoretical …

When and why does motor preparation arise in recurrent neural network models of motor control?

Marine Schimel, Ta-Chu Kao, Guillaume Hennequin

During delayed ballistic reaches, motor areas consistently display movement-specific activity patterns prior to movement onset. It is unclear why these patterns arise: while they have been proposed to seed an initial neural state from which the movement unfolds, recent experiments have uncovered …

gb, us (code pays fourni par la source)

Author response: When and why does motor preparation arise in recurrent neural network models of motor control?

Marine Schimel, Ta-Chu Kao, Guillaume Hennequin

gb, us (code pays fourni par la source)

Author response: When and why does motor preparation arise in recurrent neural network models of motor control?

Marine Schimel, Ta-Chu Kao, Guillaume Hennequin

During delayed ballistic reaches, motor areas consistently display movement-specific activity patterns prior to movement onset. It is unclear why these patterns arise: while they have been proposed to seed an initial neural state from which the movement unfolds, recent experiments have uncovered …

gb, us (code pays fourni par la source)

When and why does motor preparation arise in recurrent neural network models of motor control?

Marine Schimel, Ta-Chu Kao, Guillaume Hennequin

Summary During delayed ballistic reaches, motor areas consistently display movement-specific activity patterns prior to movement onset. It is unclear why these patterns arise: while they have been proposed to seed an initial neural state from which the movement unfolds, recent experiments have …

gb, us (code pays fourni par la source)

Kinetic features dictate sensorimotor alignment in the superior colliculus

Ana González-Rueda, Kristopher T. Jensen, Mohammadreza Noormandipour, Daniel de Malmazet et autres

Abstract The execution of goal-oriented behaviours requires a spatially coherent alignment between sensory and motor maps. The current model for sensorimotor transformation in the superior colliculus relies on the topographic mapping of static spatial receptive fields onto movement endpoints1–6. Here, to experimentally …

gb (code pays fourni par la source)