Skip to search formSkip to main contentSkip to account menu

A solution to the learning dilemma for recurrent networks of spiking neurons

The resulting learning method – called e-prop – approaches the performance of BPTT (backpropagation through time), the best known method for training recurrent neural networks in machine learning, and is biologically plausible.
View on Springer (opens in a new tab)

Biologically inspired alternatives to backpropagation through time for learning in recurrent neural nets

It is shown that an online merging of locally available information during a computation with suitable top-down learning signals in real-time provides highly capable approximations to back-propagation through time (BPTT).
View PDF on arXiv (opens in a new tab)

2022 roadmap on neuromorphic computing and engineering

The aim of this roadmap is to present a snapshot of the present state of neuromorph technology and provide an opinion on the challenges and opportunities that the future holds in the major areas of neuromorphic technology, namely materials, devices, neuromorphic circuits, neuromorph algorithms, applications, and ethics.
View on IOP Publishing (opens in a new tab)

Neuromorphic Hardware Learns to Learn

This work employs other powerful gradient-free optimization tools, such as cross-entropy methods and evolutionary strategies, in order to port the function of biological optimization processes to neuromorphic hardware and shows these optimization algorithms enable neuromorphic agents to learn very efficiently from rewards.
View on PubMed (opens in a new tab)

2021 Roadmap on Neuromorphic Computing and Engineering

This roadmap envisages the potential applications of neuromorphic materials in cutting edge technologies and focuses on the design and fabrication of artificial neural systems, which takes inspiration from biology, physics, mathematics, computer science and engineering.
View via Publisher (opens in a new tab)

One-shot learning with spiking neural networks

It is found that a corresponding model architecture, where learning signals are emitted by a separate RSNN that is optimized to facilitate fast learning, enables one-shot learning via local synaptic plasticity in RSNNs for large families of learning tasks.
View via Publisher (opens in a new tab)

Reservoirs learn to learn

It is found that this two-tiered process substantially improves the learning speed of liquid state machines for specific tasks, and this learning speed increases further if one does not train the weights of linear readouts at all, and relies instead on the internal dynamics and fading memory of the network for remembering salient information.
View PDF on arXiv (opens in a new tab)

Eligibility traces provide a data-inspired alternative to backpropagation through time

E eligibility propagation (e-prop), a new factorization of the loss gradients in RNNs that fits the framework of three factor learning rules when derived for biophysical spiking neuron models is presented.

Self-Supervised Learning Through Efference Copies

This work formally recover and extend SSL methods such as SimCLR, BYOL, and ReLIC under a common theoretical framework, i.e. Self-supervision Through Efference Copies (S-TEC), and hypothesize a testable positive from the brain’s motor outputs onto its sensory representations.
View PDF on arXiv (opens in a new tab)

CCN GAC Workshop: Issues with learning in biological recurrent neural networks

Recommendations for both theoretical and experimental neuroscientists when designing new studies that could help to bring clarity to the common assumptions about biological learning and the corresponding findings from experimental neuroscience are concluded.
By clicking accept or continuing to use the site, you agree to the terms outlined in our Privacy Policy (opens in a new tab), Terms of Service (opens in a new tab), and Dataset License (opens in a new tab)