Madison Cotteret
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Madison Cotteret has a total of 23 co-authors. Here are the top co-authors by number of shared publications:
- Hugh Greatorex
University of Groningen, Groningen, Netherlands
7 shared publications - Elisabetta Chicca
University of Groningen
6 shared publications - Michele Mastella
Neuronova, Milan, Italy
4 shared publications - Ole Richter
SynSense AG Zurich Switzerland
4 shared publications Martin Ziegler
Energy Materials and Devices, Department of Materials Science, Kiel University, Kiel, Germany
3 shared publications- Arianna Rubino
ETH Zurich, Zurich, Switzerland
2 shared publications
Latest Publications
- Open Access
- 4 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 67-74
Open-source asynchronous periphery for mixed-signal neuromorphic chips
Mixed-signal neuromorphic systems combine analog neural dynamics with event-driven asynchronous digital communication. Designing the asynchronous digital periphery for these systems is often a major practical barrier however, in part due to the lack of available reusable, silicon-verified components. This forces many research groups to re-design core components from scratch, increasing development time and risk. To address this, we present an open-source library of silicon-verified asynchronous peripheral circuits designed for mixed-signal neuromorphic processors, using the Asynchronous Circuit Toolkit (ACT) with Quasi-Delay-Insensitive (QDI) design methodology and Dual-Rail (DR) encoding. We describe the circuit templates, outline a multi-level verification methodology from behavioural simulation through post-fabrication testing, and present timing measurements from fabricated silicon.
- 4 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 67-74
Open AccessOpen-source asynchronous periphery for mixed-signal neuromorphic chips
Mixed-signal neuromorphic systems combine analog neural dynamics with event-driven asynchronous digital communication. Designing the asynchronous digital periphery for these systems is often a major practical barrier however, in part due to the lack of available reusable, silicon-verified components. This forces many research groups to re-design core components from scratch, increasing development time and risk. To address this, we present an open-source library of silicon-verified asynchronous peripheral circuits designed for mixed-signal neuromorphic processors, using the Asynchronous Circuit Toolkit (ACT) with Quasi-Delay-Insensitive (QDI) design methodology and Dual-Rail (DR) encoding. We describe the circuit templates, outline a multi-level verification methodology from behavioural simulation through post-fabrication testing, and present timing measurements from fabricated silicon.
- Open Access
- 11 July 2025
- Published by Springer Nature in Nature Communications
- Vol. 16 (1) , 1-14
A neuromorphic processor with on-chip learning for beyond-CMOS device integration
Recent advances in memory technologies, devices, and materials have shown great potential for integration into neuromorphic electronic systems. However, a significant gap remains between the development of these materials and the realization of large-scale, fully functional systems. One key challenge is determining which devices and materials are best suited for specific functions and how they can be paired with complementary metal-oxide-semiconductor circuitry. To address this, we present a mixed-signal neuromorphic architecture designed to explore the integration of on-chip learning circuits and novel two- and three-terminal devices. The chip serves as a platform to bridge the gap between silicon-based neuromorphic computation and the latest advancements in emerging devices. In this paper, we demonstrate the readiness of the architecture for device integration through comprehensive measurements and simulations. The processor provides a practical system for testing bio-inspired learning algorithms alongside emerging devices, establishing a tangible link between brain-inspired computation and cutting-edge device research.
- 11 July 2025
- Published by Springer Nature in Nature Communications
- Vol. 16 (1) , 1-14
Open AccessA neuromorphic processor with on-chip learning for beyond-CMOS device integration
Recent advances in memory technologies, devices, and materials have shown great potential for integration into neuromorphic electronic systems. However, a significant gap remains between the development of these materials and the realization of large-scale, fully functional systems. One key challenge is determining which devices and materials are best suited for specific functions and how they can be paired with complementary metal-oxide-semiconductor circuitry. To address this, we present a mixed-signal neuromorphic architecture designed to explore the integration of on-chip learning circuits and novel two- and three-terminal devices. The chip serves as a platform to bridge the gap between silicon-based neuromorphic computation and the latest advancements in emerging devices. In this paper, we demonstrate the readiness of the architecture for device integration through comprehensive measurements and simulations. The processor provides a practical system for testing bio-inspired learning algorithms alongside emerging devices, establishing a tangible link between brain-inspired computation and cutting-edge device research.
- Open Access
High-resolution spatial memory requires grid-cell-like neural codes
Continuous attractor networks (CANs) are widely used to model how the brain temporarily retains continuous behavioural variables via persistent recurrent activity, such as an animal's position in an environment. However, this memory mechanism is very sensitive to even small imperfections, such as noise or heterogeneity, which are both common in biological systems. Previous work has shown that discretising the continuum into a finite set of discrete attractor states provides robustness to these imperfections, but necessarily reduces the resolution of the represented variable, creating a dilemma between stability and resolution. We show that this stability-resolution dilemma is most severe for CANs using unimodal bump-like codes, as in traditional models. To overcome this, we investigate sparse binary distributed codes based on random feature embeddings, in which neurons have spatially-periodic receptive fields. We demonstrate theoretically and with simulations that such grid-cell-like codes enable CANs to achieve both high stability and high resolution simultaneously. The model extends to embedding arbitrary nonlinear manifolds into a CAN, such as spheres or tori, and generalises linear path integration to integration along freely-programmable on-manifold vector fields. Together, this work provides a theory of how the brain could robustly represent continuous variables with high resolution and perform flexible computations over task-relevant manifolds.
Open AccessHigh-resolution spatial memory requires grid-cell-like neural codes
Continuous attractor networks (CANs) are widely used to model how the brain temporarily retains continuous behavioural variables via persistent recurrent activity, such as an animal's position in an environment. However, this memory mechanism is very sensitive to even small imperfections, such as noise or heterogeneity, which are both common in biological systems. Previous work has shown that discretising the continuum into a finite set of discrete attractor states provides robustness to these imperfections, but necessarily reduces the resolution of the represented variable, creating a dilemma between stability and resolution. We show that this stability-resolution dilemma is most severe for CANs using unimodal bump-like codes, as in traditional models. To overcome this, we investigate sparse binary distributed codes based on random feature embeddings, in which neurons have spatially-periodic receptive fields. We demonstrate theoretically and with simulations that such grid-cell-like codes enable CANs to achieve both high stability and high resolution simultaneously. The model extends to embedding arbitrary nonlinear manifolds into a CAN, such as spheres or tori, and generalises linear path integration to integration along freely-programmable on-manifold vector fields. Together, this work provides a theory of how the brain could robustly represent continuous variables with high resolution and perform flexible computations over task-relevant manifolds.
- Open Access
- 7 February 2025
- Published by IOP Publishing in Neuromorphic Computing and Engineering
- Vol. 5 (1) , 014008
Distributed representations enable robust multi-timescale symbolic computation in neuromorphic hardware
Programming recurrent spiking neural networks (RSNNs) to robustly perform multi-timescale computation remains a difficult challenge. To address this, we describe a single-shot weight learning scheme to embed robust multi-timescale dynamics into attractor-based RSNNs, by exploiting the properties of high-dimensional distributed representations. We embed finite state machines into the RSNN dynamics by superimposing a symmetric autoassociative weight matrix and asymmetric transition terms, which are each formed by the vector binding of an input and heteroassociative outer-products between states. Our approach is validated through simulations with highly nonideal weights; an experimental closed-loop memristive hardware setup; and on Loihi 2, where it scales seamlessly to large state machines. This work introduces a scalable approach to embed robust symbolic computation through recurrent dynamics into neuromorphic hardware, without requiring parameter fine-tuning or significant platform-specific optimisation. Moreover, it demonstrates that distributed symbolic representations serve as a highly capable representation-invariant language for cognitive algorithms in neuromorphic hardware.
- 7 February 2025
- Published by IOP Publishing in Neuromorphic Computing and Engineering
- Vol. 5 (1) , 014008
Open AccessDistributed representations enable robust multi-timescale symbolic computation in neuromorphic hardware
Programming recurrent spiking neural networks (RSNNs) to robustly perform multi-timescale computation remains a difficult challenge. To address this, we describe a single-shot weight learning scheme to embed robust multi-timescale dynamics into attractor-based RSNNs, by exploiting the properties of high-dimensional distributed representations. We embed finite state machines into the RSNN dynamics by superimposing a symmetric autoassociative weight matrix and asymmetric transition terms, which are each formed by the vector binding of an input and heteroassociative outer-products between states. Our approach is validated through simulations with highly nonideal weights; an experimental closed-loop memristive hardware setup; and on Loihi 2, where it scales seamlessly to large state machines. This work introduces a scalable approach to embed robust symbolic computation through recurrent dynamics into neuromorphic hardware, without requiring parameter fine-tuning or significant platform-specific optimisation. Moreover, it demonstrates that distributed symbolic representations serve as a highly capable representation-invariant language for cognitive algorithms in neuromorphic hardware.
- Open Access
- 21 March 2024
- Published by MIT Press in Neural Computation
- Vol. 36 (4) , 549-595
Vector Symbolic Finite State Machines in Attractor Neural Networks
Hopfield attractor networks are robust distributed models of human memory, but they lack a general mechanism for effecting state-dependent attractor transitions in response to input. We propose construction rules such that an attractor network may implement an arbitrary finite state machine (FSM), where states and stimuli are represented by high-dimensional random vectors and all state transitions are enacted by the attractor network’s dynamics. Numerical simulations show the capacity of the model, in terms of the maximum size of implementable FSM, to be linear in the size of the attractor network for dense bipolar state vectors and approximately quadratic for sparse binary state vectors. We show that the model is robust to imprecise and noisy weights, and so a prime candidate for implementation with high-density but unreliable devices. By endowing attractor networks with the ability to emulate arbitrary FSMs, we propose a plausible path by which FSMs could exist as a distributed computational primitive in biological neural networks.
- 21 March 2024
- Published by MIT Press in Neural Computation
- Vol. 36 (4) , 549-595
Open AccessVector Symbolic Finite State Machines in Attractor Neural Networks
Hopfield attractor networks are robust distributed models of human memory, but they lack a general mechanism for effecting state-dependent attractor transitions in response to input. We propose construction rules such that an attractor network may implement an arbitrary finite state machine (FSM), where states and stimuli are represented by high-dimensional random vectors and all state transitions are enacted by the attractor network’s dynamics. Numerical simulations show the capacity of the model, in terms of the maximum size of implementable FSM, to be linear in the size of the attractor network for dense bipolar state vectors and approximately quadratic for sparse binary state vectors. We show that the model is robust to imprecise and noisy weights, and so a prime candidate for implementation with high-density but unreliable devices. By endowing attractor networks with the ability to emulate arbitrary FSMs, we propose a plausible path by which FSMs could exist as a distributed computational primitive in biological neural networks.
- Open Access
Vector Symbolic Finite State Machines in Attractor Neural Networks
Hopfield attractor networks are robust distributed models of human memory, but lack a general mechanism for effecting state-dependent attractor transitions in response to input. We propose construction rules such that an attractor network may implement an arbitrary finite state machine (FSM), where states and stimuli are represented by high-dimensional random vectors, and all state transitions are enacted by the attractor network's dynamics. Numerical simulations show the capacity of the model, in terms of the maximum size of implementable FSM, to be linear in the size of the attractor network for dense bipolar state vectors, and approximately quadratic for sparse binary state vectors. We show that the model is robust to imprecise and noisy weights, and so a prime candidate for implementation with high-density but unreliable devices. By endowing attractor networks with the ability to emulate arbitrary FSMs, we propose a plausible path by which FSMs could exist as a distributed computational primitive in biological neural networks.
Open AccessVector Symbolic Finite State Machines in Attractor Neural Networks
Hopfield attractor networks are robust distributed models of human memory, but lack a general mechanism for effecting state-dependent attractor transitions in response to input. We propose construction rules such that an attractor network may implement an arbitrary finite state machine (FSM), where states and stimuli are represented by high-dimensional random vectors, and all state transitions are enacted by the attractor network's dynamics. Numerical simulations show the capacity of the model, in terms of the maximum size of implementable FSM, to be linear in the size of the attractor network for dense bipolar state vectors, and approximately quadratic for sparse binary state vectors. We show that the model is robust to imprecise and noisy weights, and so a prime candidate for implementation with high-density but unreliable devices. By endowing attractor networks with the ability to emulate arbitrary FSMs, we propose a plausible path by which FSMs could exist as a distributed computational primitive in biological neural networks.
- Open Access
- 28 August 2023
- Published by Association for Computing Machinery (ACM)
- p. 1-4
Synaptic Normalisation for On-Chip Learning in Analog CMOS Spiking Neural Networks
- 28 August 2023
- Published by Association for Computing Machinery (ACM)
- p. 1-4
Open AccessSynaptic Normalisation for On-Chip Learning in Analog CMOS Spiking Neural Networks
- Open Access
- 28 August 2023
- Published by Association for Computing Machinery (ACM)
- p. 1-4
A Subthreshold Second-Order Integration Circuit for Versatile Synaptic Alpha Kernel and Trace Generation
In neuromorphic hardware the choice of synaptic kernels and trace dynamics are key for the correct system abstraction and representation of information. This paper presents a novel second-order integration circuit for the implementation of traces and kernels, the Second-order Differential Pair Integrator (SoDPI). It provides smooth alpha-kernel shaped responses to spike input in analog subthreshold complementary Metal Oxide Semiconductor (MOS) technology. Our approach utilises two Differential Pair Integrator (DPI) circuits in series to implement an effective current-mode second-order translinear low-pass filter. Theoretical analysis and experimental measurements demonstrate the improved reliability of this design, which offers a promising approach for modelling biological synaptic and neural responses in neuromorphic hardware, as well as improving the stability of integrated on-chip learning systems.
- 28 August 2023
- Published by Association for Computing Machinery (ACM)
- p. 1-4
Open AccessA Subthreshold Second-Order Integration Circuit for Versatile Synaptic Alpha Kernel and Trace Generation
In neuromorphic hardware the choice of synaptic kernels and trace dynamics are key for the correct system abstraction and representation of information. This paper presents a novel second-order integration circuit for the implementation of traces and kernels, the Second-order Differential Pair Integrator (SoDPI). It provides smooth alpha-kernel shaped responses to spike input in analog subthreshold complementary Metal Oxide Semiconductor (MOS) technology. Our approach utilises two Differential Pair Integrator (DPI) circuits in series to implement an effective current-mode second-order translinear low-pass filter. Theoretical analysis and experimental measurements demonstrate the improved reliability of this design, which offers a promising approach for modelling biological synaptic and neural responses in neuromorphic hardware, as well as improving the stability of integrated on-chip learning systems.