Learning automata based energy-efficient AI hardware design for IoT applications

Abstract

Energy efficiency continues to be the core design challenge for artificial intelligence (AI) hardware designers. In this paper, we propose a new AI hardware architecture targeting Internet of Things applications. The architecture is founded on the principle of learning automata, defined using propositional logic. The logic-based underpinning enables low-energy footprints as well as high learning accuracy during training and inference, which are crucial requirements for efficient AI with long operating life. We present the first insights into this new architecture in the form of a custom-designed integrated circuit for pervasive applications. Fundamental to this circuit is systematic encoding of binarized input data fed into maximally parallel logic blocks. The allocation of these blocks is optimized through a design exploration and automation flow using field programmable gate array-based fast prototypes and software simulations. The design flow allows for an expedited hyperparameter search for meeting the conflicting requirements of energy frugality and high accuracy. Extensive validations on the hardware implementation of the new architecture using single- and multi-class machine learning datasets show potential for significantly lower energy than the existing AI hardware architectures. In addition, we demonstrate test accuracy and robustness matching the software implementation, outperforming other state-of-the-art machine learning algorithms. This article is part of the theme issue ‘Advanced electromagnetic non-destructive evaluation and smart monitoring’.

Low-Power Audio Keyword Spotting using Tsetlin Machines

“…These Boolean 119 features are also referred to as literals:X and X. Current research has shown that 120 significance-driven Booleanization of features for the Tsetlin Machine is vital in control-121 ling the Tsetlin Machine size and processing requirements [18]. Increasing the number of 122 features will increase the number of TA and increase computations for the clause module 123 and subsequently the energy spent in incrementing and decrementing states in the 124 feedback module.…”

“…Audio data streams are always subject to redundancies in the channel that formal-156 ize as nonvocal noise, background noise and silence [20,21]. Therefore the challenge [18].…”

“…the inter-class features start to become increasingly correlated. Correlated 368 inter-class features will lead to class overlap and degrade TM performance [18]. Through To mitigate against the effect on performance of increasing keywords, there are two 372 methods available: Firstly to adjust the Tsetlin Machines hyperparameters to enable 373 more events triggered (see Figure 2).…”

“…The major factor that has impacted the performance of the KWS is the capacity 443 of the TM which is determined by the number of clauses per class. The higher the 444 number clauses, the higher the overall classification accuracy [18]. Yet, the resource 445 usage will increase linearly along with the energy consumption and memory footprint.…”

“…It is well defined that the number of input features to the TM is one of the major 272 factors that affect its resource usage[17][18][19]. Increased raw input features means more273 Booleans are required to represent them and thus the number of Tsetlin Automaton (TA) 274…”