Zobrazeno 1 - 10
of 52
pro vyhledávání: '"Subramoney, Anand"'
Autor:
Schöne, Mark, Bhisikar, Yash, Bania, Karan, Nazeer, Khaleelulla Khan, Mayr, Christian, Subramoney, Anand, Kappel, David
Handling sparse and unstructured geometric data, such as point clouds or event-based vision, is a pressing challenge in the field of machine vision. Recently, sequence models such as Transformers and state-space models entered the domain of geometric
Externí odkaz:
http://arxiv.org/abs/2411.12603
Autor:
Sushma, Neeraj Mohan, Tian, Yudou, Mestha, Harshvardhan, Colombo, Nicolo, Kappel, David, Subramoney, Anand
Deep state-space models (Deep SSMs) have shown capabilities for in-context learning on autoregressive tasks, similar to transformers. However, the architectural requirements and mechanisms enabling this in recurrent networks remain unclear. This stud
Externí odkaz:
http://arxiv.org/abs/2410.11687
Autor:
Fokam, Cabrel Teguemne, Nazeer, Khaleelulla Khan, König, Lukas, Kappel, David, Subramoney, Anand
The increasing size of deep learning models has created the need for more efficient alternatives to the standard error backpropagation algorithm, that make better use of asynchronous, parallel and distributed computing. One major shortcoming of backp
Externí odkaz:
http://arxiv.org/abs/2410.05985
Hierarchical model-based reinforcement learning (HMBRL) aims to combine the benefits of better sample efficiency of model based reinforcement learning (MBRL) with the abstraction capability of hierarchical reinforcement learning (HRL) to solve comple
Externí odkaz:
http://arxiv.org/abs/2406.00483
Autor:
Mukherji, Rishav, Schöne, Mark, Nazeer, Khaleelulla Khan, Mayr, Christian, Kappel, David, Subramoney, Anand
Activity and parameter sparsity are two standard methods of making neural networks computationally more efficient. Event-based architectures such as spiking neural networks (SNNs) naturally exhibit activity sparsity, and many methods exist to sparsif
Externí odkaz:
http://arxiv.org/abs/2405.00433
Autor:
Schöne, Mark, Sushma, Neeraj Mohan, Zhuge, Jingyue, Mayr, Christian, Subramoney, Anand, Kappel, David
Event-based sensors are well suited for real-time processing due to their fast response times and encoding of the sensory data as successive temporal differences. These and other valuable properties, such as a high dynamic range, are suppressed when
Externí odkaz:
http://arxiv.org/abs/2404.18508
Autor:
Nazeer, Khaleelulla Khan, Schöne, Mark, Mukherji, Rishav, Vogginger, Bernhard, Mayr, Christian, Kappel, David, Subramoney, Anand
As large language models continue to scale in size rapidly, so too does the computational power required to run them. Event-based networks on neuromorphic devices offer a potential way to reduce energy consumption for inference significantly. However
Externí odkaz:
http://arxiv.org/abs/2312.09084
Artificial neural networks open up unprecedented machine learning capabilities at the cost of ever growing computational requirements. Sparsifying the parameters, often achieved through weight pruning, has been identified as a powerful technique to c
Externí odkaz:
http://arxiv.org/abs/2311.07625
Biological evidence suggests that adaptation of synaptic delays on short to medium timescales plays an important role in learning in the brain. Inspired by biology, we explore the feasibility and power of using synaptic delays to solve challenging ta
Externí odkaz:
http://arxiv.org/abs/2306.06237
Autor:
Kappel, David, Nazeer, Khaleelulla Khan, Fokam, Cabrel Teguemne, Mayr, Christian, Subramoney, Anand
The ubiquitous backpropagation algorithm requires sequential updates through the network introducing a locking problem. In addition, back-propagation relies on the transpose of forward weight matrices to compute updates, introducing a weight transpor
Externí odkaz:
http://arxiv.org/abs/2305.14974