Align, then memorise: the dynamics of learning with feedback alignment

Chat with Graph Search

Ask any question about EPFL courses, lectures, exercises, research, news, etc. or try the example questions below.

DISCLAIMER: The Graph Chatbot is not programmed to provide explicit or categorical answers to your questions. Rather, it transforms your questions into API requests that are distributed across the various IT services officially administered by EPFL. Its purpose is solely to collect and recommend relevant references to content that you can explore to help you answer your questions.

In this thesis, we study how combinatorial structures and symmetries emerge from analytic properties of graph polynomials and neural networks. Graph polynomials encode combinatorial aspects of the underlying graph in their coefficients and are remarkably u ...

Electrochemical gas sensors (EGSs) have been used to measure the surface distributions and vertical profiles of trace gases in the wintertime Arctic boundary layer during the Alaskan Layered Pollution and Chemical Analysis (ALPACA) field experiment in Fair ...

Restoring natural sensation via neuroprosthetics relies on the possibility of encoding complex and nuanced information. For example, an ideal brain–machine interface with sensory feedback would provide the user with sensation about movement, pressure, curv ...

For high-dimensional Gaussian data, we investigate theoretically how the features of a two-layer neural network adapt to the structure of the target function through a few large batch gradient descent steps, leading to an improvement in the approximation c ...

Virtual reality (VR) technology revitalises rehabilitation training by creating rich, interactive virtual rehabilitation scenes and tasks that deeply engage patients. Robotics with immersive VR environments have the potential to significantly enhance the s ...

Forthcoming imaging surveys will increase the number of known galaxy-scale strong lenses by several orders of magnitude. For this to happen, images of billions of galaxies will have to be inspected to identify potential candidates. In this context, deep-le ...

Understanding the brain requires studying its multiscale interactions from molecules to networks. The increasing availability of large-scale datasets detailing brain circuit composition, connectivity, and activity is transforming neuroscience. However, int ...

Transformer-based architectures are the model of choice for natural language understanding, but they come at a significant cost, as they have quadratic complexity in the input length, require a lot of training data, and can be difficult to tune. In the pur ...

Inferring the source of a diffusion in a large network of agents is a difficult but feasible task, if a few agents act as sensors revealing the time at which they got hit by the diffusion. One of the main limitations of current source identification algori ...

Relating behavior to brain activity in animals is a fundamental goal in neuroscience, with practical applications in building robust brain-machine interfaces. However, the domain gap between individuals is a major issue that prevents the training of genera ...

Autoencoders are the simplest neural network for unsupervised learning, and thus an ideal framework for studying feature learning. While a detailed understanding of the dynamics of linear autoencoders has recently been obtained, the study of non-linear aut ...

We present an approach to quantifying both aleatoric and epistemic uncertainty for deep neural networks in image classification, based on generative adversarial networks (GANs). While most works in the literature that use GANs to generate out-of-distributi ...

We present a method to maintain the subjective perception of volume of audio signals and, at the same time, reduce their absolute peak value. We focus on achieving this without compromising the perceived audio quality. This is specially useful, for example ...

In this thesis, we study the 3 challenges described above. First, we study different reconstruction techniques and assess the fidelity of each reconstruction results by means of structured illumination and phase conjugation. By reconstructing the 3D refrac ...

Gain-cell-embedded DRAM (GC-eDRAM) is an attractive alternative to traditional 6T SRAM, as it offers higher density, lower leakage power, and two-ported functionality. However, its refresh requirement also results in power consumption and memory access lim ...

Physics-based representations constructed using only atomic positions and nuclear charges (also known as quantum machine learning, QML) allow for the reliable and efficient inference of molecular properties from training data. Chemistry is a science rooted ...

We investigate how the training curve of isotropic kernel methods depends on the symmetry of the task to be learned, in several settings. (i) We consider a regression task, where the target function is a Gaussian random field that depends only on d(paralle ...

Modern information technologies and human-centric communication systems employ advanced content representations for richer portrayals of the real world. The newly adopted imaging modalities offer additional information cues and permit the depiction of real ...

Align, then memorise: the dynamics of learning with feedback alignment

Chat with Graph Search

Combinatorial Aspects of Continuous Optimization: Graph Polynomials and Neural Networks

Surface distributions and vertical profiles of trace gases (CO, O3, NO, NO2) in the Arctic wintertime boundary layer using low-cost sensors during ALPACA-2022

Optimization frameworks for bespoke sensory encoding in neuroprosthetics

How Two-Layer Neural Networks Learn, One (Giant) Step at a Time

Rehabilitation exoskeleton system with bidirectional virtual reality feedback training strategy

Euclid preparation XXXIII. Characterization of convolutional neural networks for the identification of galaxy-galaxy strong-lensing events

Large-Scale Mechanistic Models of Brain Circuits with Biophysically and Morphologically Detailed Neurons

How can LLMs transform the robotic design process?

HyperMixer: An MLP-based Low Cost Alternative to Transformers

FM -Directories: Extending the Burrows-Wheeler Transform for String Labeled Vertex Graphs of (Almost) Arbitrary Topology

Source identification via contact tracing in the presence of asymptomatic patients

Overcoming the Domain Gap in Neural Action Representations

The dynamics of representation learning in shallow, non-linear autoencoders

UQGAN: A Unified Model for Uncertainty Quantification of Deep Classifiers trained via Conditional GANs

Increasing Loudness In Audio Signals: A Perceptually Motivated Approach To Preserve Audio Quality

Advanced Techniques in Optical Diffraction Tomography

64-kB 65-nm GC-eDRAM With Half-Select Support and Parallel Refresh Technique

Physics-based representations for machine learning properties of chemical reactions

How isotropic kernels perform on simple invariants

Perceptual Quality of Point Clouds with application to Compression