Low-Power Audio Keyword Spotting using Tsetlin Machines

Lei, Jie; Rahman, Tousif; Shafik, Rishad; Yakovlev, Alex; Granmo, Ole‐Christoffer; Kawsar, Fahim; Mathur, and

doi:10.20944/preprints202101.0621.v1

Cited by 7 publications

(4 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, some energy-efficient Machine Learning alternatives have been proposed. Among those is the work of [13], presenting a KWS system based on a Tsetlin Machine, a Finite State Machine exploiting propositional logic to perform classification, while the work of [14] investigates an approach based on Support Vector Machines (SVMs). The currently best performing models are compared in [15], typically referring to 12 classes classification problems on the common benchmark represented by the Google Speech Commands dataset [3].…”

Section: Related Workmentioning

confidence: 99%

Target-Aware Neural Architecture Search and Deployment for Keyword Spotting

et al. 2022

View full text Add to dashboard Cite

Keyword spotting (KWS) utilities have become increasingly popular on a wide range of mobile and home devices, representing a prolific application field for Convolutional Neural Networks (CNNs), which are commonly exploited to perform keyword classification. Addressing the challenges of targeting such resource-constrained platforms, requires a careful definition of the CNN architecture and the overall system implementation. These reasons have led to a growing need for design and optimization flows, able to intrinsically take into account the system's performance when ported on the target platform. In this work, we present a design methodology based on Neural Architecture Search, exploited to combine the exploration of the optimal network topology, the audio pre-processing scheme, and the data quantization policy. The proposed design flow includes target-awareness in the exploration loop, comparing the different design alternatives according to a model-based pre-evaluation of metrics like execution latency, memory footprint, and energy consumption, evaluated considering the application's execution on the target processing platform. We have tested our design flow to obtain target-specific CNNs for a resource-constrained commercial platform, the ST SensorTile. Considering two different application scenarios, enabling the comparison with the state-of-the-art of efficient CNN-based models for KWS, we have obtained up to a 1.8% accuracy improvement and a 40% footprint reduction in the most favorable case.INDEX TERMS Keyword spotting, neural networks, neural architecture search.

show abstract

Section: Related Workmentioning

confidence: 99%

Target-Aware Neural Architecture Search and Deployment for Keyword Spotting

et al. 2022

View full text Add to dashboard Cite

show abstract

“…Recent works show Tsetlin machines to have successfully addressed several machine learning tasks, including natural language understanding (Berge et al, 2019;Bhattarai et al, 2021;Saha et al, 2020;Yadav et al, 2021aYadav et al, , 2021b, speech understanding (Lei et al, 2021), image analysis (Granmo et al, 2019), classification (Abeyrathna et al, 2021), and regression (Darshana Abeyrathna et al, 2020). While the performance showed by Tsetlin machines in such areas has been comparable to state-of-the-art machine learning techniques, the method has also been shown to have a smaller memory footprint and faster inference in reported cases than more traditional neural network-based models (Granmo et al, 2019;Wheeldon et al, 2020).…”

Section: Tsetlin Machinesmentioning

confidence: 99%

Using Tsetlin Machine to discover interpretable rules in natural language processing applications

2021

View full text Add to dashboard Cite

Tsetlin Machines (TM) use finite state machines for learning and propositional logic to represent patterns. The resulting pattern recognition approach captures information in the form of conjunctive clauses, thus facilitating human interpretation. In this work, we propose a TM-based approach to three common natural language processing (NLP) tasks, namely, sentiment analysis, semantic relation categorization and identifying entities in multi-turn dialogues. By performing frequent itemset mining on the TM-produced patterns, we show that we can obtain a global and a local interpretation of the learning, one that mimics existing rule-sets or lexicons. Further, we also establish that our TM based approach does not compromise on accuracy in the quest for interpretability, via comparison with some widely used machine learning techniques. Finally, we introduce the idea of a relational TM, which uses a logicbased framework to further extend the interpretability.

show abstract

“…Several researchers have lately explored various TMbased natural language processing models, including text classification [4,21], novelty detection [5], semantic relation analysis [22], and aspect-based sentiment analysis [27], using conjunctive clauses to capture textual patterns. Other application areas are network attack detection [11], keyword spotting [12], biomedical systems design [16], and game playing [9]. Further, the vanilla TM has been significantly extended by weighted clauses [13], regression architectures [2], and the elimination of hyperparameters [15].…”

Section: Introductionmentioning

confidence: 99%

Off-policy and on-policy reinforcement learning with the Tsetlin machine

Gorji

Granmo

2023

Appl Intell

View full text Add to dashboard Cite

The Tsetlin Machine is a recent supervised learning algorithm that has obtained competitive accuracy- and resource usage results across several benchmarks. It has been used for convolution, classification, and regression, producing interpretable rules in propositional logic. In this paper, we introduce the first framework for reinforcement learning based on the Tsetlin Machine. Our framework integrates the value iteration algorithm with the regression Tsetlin Machine as the value function approximator. To obtain accurate off-policy state-value estimation, we propose a modified Tsetlin Machine feedback mechanism that adapts to the dynamic nature of value iteration. In particular, we show that the Tsetlin Machine is able to unlearn and recover from the misleading experiences that often occur at the beginning of training. A key challenge that we address is mapping the intrinsically continuous nature of state-value learning to the propositional Tsetlin Machine architecture, leveraging probabilistic updates. While accurate off-policy, this mechanism learns significantly slower than neural networks on-policy. However, by introducing multi-step temporal-difference learning in combination with high-frequency propositional logic patterns, we are able to close the performance gap. Several gridworld instances document that our framework can outperform comparable neural network models, despite being based on simple one-level AND-rules in propositional logic. Finally, we propose how the class of models learnt by our Tsetlin Machine for the gridworld problem can be translated into a more understandable graph structure. The graph structure captures the state-value function approximation and the corresponding policy found by the Tsetlin Machine.

show abstract

Low-Power Audio Keyword Spotting using Tsetlin Machines

Cited by 7 publications

References 12 publications

Target-Aware Neural Architecture Search and Deployment for Keyword Spotting

Target-Aware Neural Architecture Search and Deployment for Keyword Spotting

Using Tsetlin Machine to discover interpretable rules in natural language processing applications

Off-policy and on-policy reinforcement learning with the Tsetlin machine

Contact Info

Product

Resources

About