paper-with-me

홈 › Papers

Improving the Robustness of Neural Multiplication Units with Reversible Stochasticity

2022-11-10 · Bhumika Mistry, Katayoun Farrahi, Jonathon Hare

Multilayer Perceptrons struggle to learn certain simple arithmetic tasks. Specialist neural modules for arithmetic can outperform classical architectures with gains in extrapolation, interpretability and convergence speeds, but are highly sensitive to the training range. In this paper, we show that Neural Multiplication Units (NMUs) are unable to reliably learn tasks as simple as multiplying two inputs when given different training ranges. Causes of failure are linked to inductive and input biases which encourage convergence to solutions in undesirable optima. A solution, the stochastic NMU (sNMU), is proposed to apply reversible stochasticity, encouraging avoidance of such optima whilst converging to the true solution. Empirically, we show that stochasticity provides improved robustness with the potential to improve learned representations of upstream networks for numerical and image tasks.

📄 PDF Abstract BibTeX arXiv:2211.05624

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Energy-Efficient Supervised Learning with a Binary Stochastic Forward-Forward Algorithm

2025-07-09 · Risi Jaiswal, Supriyo Datta, Joseph G. Makin arxiv

Reducing energy consumption has become a pressing need for modern machine learning, which has achieved many of its most impressive results by scaling to larger and more energy-consumptive neural networks. Unfortunately, …

Local Competition and Stochasticity for Adversarial Robustness in Deep Learning

2021-01-04 · Konstantinos P. Panousis, Sotirios Chatzis, Antonios Alexos, Sergios Theodoridis

This work addresses adversarial robustness in deep learning by considering deep networks with stochastic local winner-takes-all (LWTA) activations. This type of network units result in sparse representations from each mo…

Adversarial AttackAdversarial RobustnessDeep Learning

A Computational Model for Tensor Core Units

2019-08-19 · Rezaul Chowdhury, Francesco Silvestri, Flavio Vella

To respond to the need of efficient training and inference of deep neural networks, a plethora of domain-specific hardware architectures have been introduced, such as Google Tensor Processing Units and NVIDIA Tensor Core…

model

Incorporating compositional heterogeneity into Lie Markov models for phylogenetic inference

2020-07-17

Phylogenetics uses alignments of molecular sequence data to learn about evolutionary trees. Substitutions in sequences are modelled through a continuous-time Markov process, characterised by an instantaneous rate matrix,…

Neural Arithmetic Units

2020-01-14 · ICLR 2020 1 · Andreas Madsen, Alexander Rosenberg Johansen

Neural networks can approximate complex functions, but they struggle to perform exact arithmetic operations over real numbers. The lack of inductive bias for arithmetic operations leaves neural networks without the under…

Inductive Bias