Abstract:Wildlife conservation using continuous monitoring of environmental factors and biomedical classification, which generate a vast amount of sensor data, is a challenge due to limited bandwidth in the case of remote monitoring. It becomes critical to have classification where data is generated, and only classified data is used for monitoring. We present a novel multiplierless framework for in-filter acoustic classification using Margin Propagation (MP) approximation used in low-power edge devices deployable in remote areas with limited connectivity. The entire design of this classification framework is based on template-based kernel machine, which include feature extraction and inference, and uses basic primitives like addition/subtraction, shift, and comparator operations, for hardware implementation. Unlike full precision training methods for traditional classification, we use MP-based approximation for training, including backpropagation mitigating approximation errors. The proposed framework is general enough for acoustic classification. However, we demonstrate the hardware friendliness of this framework by implementing a parallel Finite Impulse Response (FIR) filter bank in a kernel machine classifier optimized for a Field Programmable Gate Array (FPGA). The FIR filter acts as the feature extractor and non-linear kernel for the kernel machine implemented using MP approximation and a downsampling method to reduce the order of the filters. The FPGA implementation on Spartan 7 shows that the MP-approximated in-filter kernel machine is more efficient than traditional classification frameworks with just less than 1K slices.

What problem does this paper attempt to address?

### What problem does this paper attempt to solve? This paper aims to solve the challenges faced due to limited bandwidth in remote monitoring environments (such as wildlife conservation and biomedical classification). Specifically, the paper focuses on how to perform classification at the data - generating location and use only the classified data for monitoring, in order to reduce the bandwidth required for transmitting a large amount of sensor data. To solve this problem, the paper proposes a new **multiplier - free in - filter - computation framework** for acoustic classification in low - power edge devices. This framework is based on the **Margin Propagation (MP) approximation** method and uses the template - based kernel machine model for feature extraction and inference. Its main features include: 1. **Multiplier - free design**: The entire classification framework uses only basic operations (such as addition/subtraction, shift, and comparison operations), thus achieving hardware - friendly and efficient computation. 2. **Combination of feature extraction and kernel function**: Combines feature extraction and non - linear kernel functions into an efficient computing system. 3. **Scalability**: According to the Internet of Things hardware constraints, users can customize memory usage. 4. **Integrated training**: Training is carried out through the MP approximation method, reducing the approximation error introduced during the filtering and classification processes. 5. **High clock frequency**: Since basic operation units (without multipliers) are used, this framework can operate at a higher clock frequency (for example, 166 MHz). In addition, the paper shows the implementation of this framework on FPGA and verifies its potential applications in identifying wildlife sounds or sounds that may indicate poaching or illegal logging through an environmental sound data set. ### Key formulas - Decision function: \[ f(x)=w^{T}K + b \] where \(K\) is the kernel function. - Decision function in the MP domain: \[ f_{\text{MP}}(x)=z_{+}-z_{-} \] where: \[ z_{+}=\text{MP}([w_{+}+K_{+}, w_{-}+K_{-}, b_{+}],\gamma_{1}) \] \[ z_{-}=\text{MP}([w_{+}+K_{-}, w_{-}+K_{+}, b_{-}],\gamma_{1}) \] \(\gamma_{1}\) is a hyperparameter obtained through gamma annealing learning, \(K_{+}=K\) and \(K_{-}=-K\). - Normalized output: \[ z = \text{MP}([z_{+}, z_{-}],\gamma_{n}) \] where \(\gamma_{n}=1\). - Output in differential form: \[ p=p_{+}-p_{-} \] where \(p_{+}=[z_{+}-z]^{+}\), \(p_{-}=[z_{-}-z]^{+}\). Through these methods, the paper provides an efficient and energy - saving acoustic classification solution, especially suitable for resource - constrained Internet of Things devices.

Multiplierless In-filter Computing for tinyML Platforms

Multiplierless MP-Kernel Machine for Energy-Efficient Edge Devices

Optimally Approximated and Unbiased Floating-Point Multiplier with Runtime Configurability

In-filter Computing For Designing Ultra-light Acoustic Pattern Recognizers

A Reconfigurable Multiplier for Signed Multiplications with Asymmetric Bit-Widths.

PAM: A Piecewise-Linearly-Approximated Floating-Point Multiplier with Unbiasedness and Configurability

A Novel Multiplierless Hardware Implementation Method for Adaptive Filter Coefficients.

ML-MCU: A Framework to Train ML Classifiers on MCU-based IoT Edge Devices

TinyM$^2$Net: A Flexible System Algorithm Co-designed Multimodal Learning Framework for Tiny Devices

An 8-bit Single Perceptron Processing Unit for Tiny Machine Learning Applications

A Multiplier Structure Based on A Novel Real-Time Csd Recoding

A Deep Learning Inference Scheme Based on Pipelined Matrix Multiplication Acceleration Design and Non-uniform Quantization

CFU Playground: Full-Stack Open-Source Framework for Tiny Machine Learning (tinyML) Acceleration on FPGAs

Towards Vision Mixture of Experts for Wildlife Monitoring on the Edge

Tiny Classifier Circuits: Evolving Accelerators for Tabular Data

TinyChirp: Bird Song Recognition Using TinyML Models on Low-power Wireless Acoustic Sensors

TinyM$^2$Net-V3: Memory-Aware Compressed Multimodal Deep Neural Networks for Sustainable Edge Deployment

Accelerating TinyML Inference on Microcontrollers through Approximate Kernels

SMaLL: A Software Framework for portable Machine Learning Libraries

Decoupled Access-Execute enabled DVFS for tinyML deployments on STM32 microcontrollers

FPGA Based Efficient Multiplier for Image Processing Applications Using Recursive Error Free Mitchell Log Multiplier and KOM Architecture