Dynamic Programming with Spiking Neural Computing

Aimone, James B.; Parekh, Ojas; Phillips, Cynthia A.; Pınar, Ali; Severa, William; Xu, Haiping

doi:10.1145/3354265.3354285

Cited by 22 publications

(14 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…When implemented on neuromorphic architectures, these algorithms promise speed and efficiency gains by exploiting fine-grain parallelism and event-based computation. Examples include computational primitives, such as sorting, max, min, and median operations [70], a wide range of graph algorithms [71]- [74], NP-complete/hard problems, such as constraint satisfaction [75], boolean satisfiability [76], dynamic programming [77], and quadratic unconstrained binary optimization [78], [79], and novel Turing-complete computational frameworks, such as Stick [80] and SN P [81].…”

Section: C O M P U T I N G W I T H T I M Ementioning

confidence: 99%

“…The shortest path search is a foundational graph algorithm, and our formulation here is representative of many similar SNN algorithms proposed in recent years that support a much broader range of graph computations [72]- [74], [77]. Loihi's encouraging quantitative results suggest that similar order-of-magnitude gains may be realized as this domain is further developed and mapped to neuromorphic hardware.…”

Section: Loihi Search Times From the Main Plot We Refer To Graph Search As Task 12 In Section VIIImentioning

confidence: 99%

See 1 more Smart Citation

Advancing Neuromorphic Computing With Loihi: A Survey of Results and Outlook

et al. 2021

View full text Add to dashboard Cite

Deep artificial neural networks apply principles of the brain's information processing that led to breakthroughs in machine learning spanning many problem domains. Neuromorphic computing aims to take this a step further to chips more directly inspired by the form and function of biological neural circuits, so they can process new knowledge, adapt, behave, and learn in real time at low power levels. Despite several decades of research, until recently, very few published results have shown that today's neuromorphic chips can demonstrate quantitative computational value. This is now changing with the advent of Intel's Loihi, a neuromorphic research processor designed to support a broad range of spiking neural networks with sufficient scale, performance, and features to deliver competitive results compared to state-of-the-art contemporary computing architectures. This survey reviews results that are obtained to date with Loihi across the major algorithmic domains under study, including deep learning approaches and novel approaches that aim to more directly harness the key features of spike-based neuromorphic hardware. While conventional feedforward deep neural networks show modest if any benefit on Loihi, more brain-inspired networks using recurrence, precise spike-timing relationships, synaptic plasticity, stochasticity, and sparsity perform certain computation with orders of magnitude lower latency and energy compared to state-of-the-art conventional approaches. These compelling neuromorphic networks solve a diverse range of problems representative of brain-like computation, such as event-based

show abstract

Section: C O M P U T I N G W I T H T I M Ementioning

confidence: 99%

Section: Loihi Search Times From the Main Plot We Refer To Graph Search As Task 12 In Section VIIImentioning

confidence: 99%

Advancing Neuromorphic Computing With Loihi: A Survey of Results and Outlook

et al. 2021

View full text Add to dashboard Cite

show abstract

“…Similar approaches and models have been investigated earlier, especially in the field of neuromorphic computing. For example, in [54][55][56][57][58] graphs are modeled using neurons and synapses, and computations are performed by exciting specific neurons which induces propagation of current in the graph and observing the spiking behavior. Although some models are more general than the one presented here and allow for solving more complex problems like dynamic programs [55], enumeration problems [57] or the longest shortest path problem [58], we are not aware of any model explicitly discussing the biological plausibility.…”

Section: Discussionmentioning

confidence: 99%

“…For example, in [54][55][56][57][58] graphs are modeled using neurons and synapses, and computations are performed by exciting specific neurons which induces propagation of current in the graph and observing the spiking behavior. Although some models are more general than the one presented here and allow for solving more complex problems like dynamic programs [55], enumeration problems [57] or the longest shortest path problem [58], we are not aware of any model explicitly discussing the biological plausibility. In fact, most of these approaches are far from being biologically plausible as they e. g. require additional artificial memory [55] or a preprocessing step that changes the graph depending on the input data [58].…”

Section: Discussionmentioning

confidence: 99%

Can the Brain Use Waves to Solve Planning Problems?

Powell

Winkel

Hopp

et al. 2021

Preprint

View full text Add to dashboard Cite

A variety of behaviors like spatial navigation or bodily motion can be formulated as graph traversal problems through cognitive maps. We present a neural network model which can solve such tasks and is compatible with a broad range of empirical findings about the mammalian neocortex and hippocampus. The neurons and synaptic connections in the model represent structures that can result from self-organization into a cognitive map via Hebbian learning, i.e. into a graph in which each neuron represents a point of some abstract task-relevant manifold and the recurrent connections encode a distance metric on the manifold. Graph traversal problems are solved by wave-like activation patterns which travel through the recurrent network and guide a localized peak of activity onto a path from some starting position to a target state.

show abstract

“…We show that in this way, we can-on a grid-identically reproduce BFS, but also Dijkstra's algorithm as well as TD(0), known from reinforcement learning [39]. Some approaches exist for neural implementation of Dijkstra's algorithm using the Hopfield network [45] or spiking neural network [46], and however, we are not aware of other neural network architecture, which could emulate different classical algorithms using the same network architecture. In fact, our approach relates to the approach that utilizes distance transforms to generate gradient maps and perform path search [47].…”

Section: B Contributionsmentioning

confidence: 98%

Finding Optimal Paths Using Networks Without Learning—Unifying Classical Approaches

Kulvičius¹,

Herzog²,

Tamošiūnaitė³

et al. 2022

IEEE Trans. Neural Netw. Learning Syst.

View full text Add to dashboard Cite

Trajectory or path planning is a fundamental issue in a wide variety of applications. In this article, we show that it is possible to solve path planning on a maze for multiple start point and endpoint highly efficiently with a novel configuration of multilayer networks that use only weighted pooling operations, for which no network training is needed. These networks create solutions, which are identical to those from classical algorithms such as breadth-first search (BFS), Dijkstra's algorithm, or TD(0). Different from competing approaches, very large mazes containing almost one billion nodes with dense obstacle configuration and several thousand importance-weighted path endpoints can this way be solved quickly in a single pass on parallel hardware.

show abstract

Dynamic Programming with Spiking Neural Computing

Cited by 22 publications

References 22 publications

Advancing Neuromorphic Computing With Loihi: A Survey of Results and Outlook

Advancing Neuromorphic Computing With Loihi: A Survey of Results and Outlook

Can the Brain Use Waves to Solve Planning Problems?

Finding Optimal Paths Using Networks Without Learning—Unifying Classical Approaches

Contact Info

Product

Resources

About