Fine-grained Visual Classification (FGVC) aims to distinguish object classes belonging to the same category, e.g., different bird species or models of vehicles. The task is more challenging than ordinary image classification due to the subtle inter-class differences. Recent works proposed deep learning models based on the vision transformer (ViT) architecture with its self-attention mechanism to locate important regions of the objects and derive global information. However, deploying them on resource-restricted internet of things (IoT) devices is challenging due to their intensive computational cost and memory footprint. Energy and power consumption varies in different IoT devices. To improve their inference efficiency, previous approaches require manually designing the model architecture and training a separate model for each computational budget. In this work, we propose Token Adaptive Vision Transformer (TAVT) that dynamically drops out tokens and can be used for various inference scenarios across many IoT devices after training the model once. Our adaptive model can switch among different token drop configurations at run time, providing instant accuracy-efficiency trade-offs. We train a vision transformer with a progressive token pruning scheme, eliminating a large number of redundant tokens in the later layers. We then conduct a multi-objective evolutionary search with the overall number of floating point operations (FLOPs) as its efficiency constraint that could be translated to energy consumption and power to find the token pruning schemes that maximize accuracy and efficiency under various computational budgets. Empirical results show that our proposed TAVT dramatically speeds up the GPU inference latency by up to 10× and reduces memory requirements and FLOPs by up to 5.5 × and 13 × respectively while achieving competitive accuracy compared to prior ViT-based state-of-the-art approaches.

While pre-trained language models like BERT have achieved impressive results on various natural language processing tasks, deploying them on resource-restricted devices is challenging due to their intensive computational cost and memory footprint. Previous approaches mainly focused on training smaller versions of a BERT model with competitive accuracy under limited computational resources. In this paper, we extend Length Adaptive Transformer and propose to design Token and Head Adaptive Transformer, which can compress and accelerate various BERT-based models via simple fine-tuning. We train a transformer with a progressive token and head pruning scheme, eliminating a large number of redundant tokens and attention heads in the later layers. Then, we conduct a multi-objective evolutionary search with the overall number of floating point operations (FLOPs) as its efficiency constraint to find joint token and head pruning strategies that maximize accuracy and efficiency under various computational budgets. Empirical studies show that a large portion of tokens and attention heads could be pruned while achieving superior performance compared to the baseline BERT-based models and Length Adaptive Transformers in various downstream NLP tasks. MobileBERT trained with our joint token and head pruning scheme achieves a GLUE score of 83.0, which is 1.4 higher than Length Adaptive Transformer and 2.9 higher than the original model.

Although numerous advances have been made in instrumental odor detection systems, these still cannot match the efficient sampling, odor discrimination, agile mobility and the olfactory acuity of odor detection dogs. A limiting step in using dogs to detect odors is the subjectivity of the translation of odor information processed by the dog to its handler. We present our preliminary efforts towards a wireless wearable system for continuous auscultation of respiratory behavior by recording internal sounds at the neck and chest by means of a commercially available electronic stethoscope to provide objective decision support for handlers. We have identified discrete features of sniffing and panting in the time domain and utilize event duration, event rate, event mean energy, and the number of consecutive events in a row to build a decision tree classifier. Since feature extraction requires segmentation of individual sniffing and panting events, we developed an adaptive method using short-time energy contour and an adaptive threshold. The performance of the system was evaluated on recordings from a Greyhound and a Labrador Retriever and achieved high classification accuracies.

Training animals is a process that requires a significant investment of time and energy on the part of the trainer. One of the most basic training tasks is to train dogs to perform postures on cue. While it might be easy for a human trainer to see when an animal has performed the desired posture, it is much more difficult for a computer to determine this. Most work in this area uses accelerometer and/or gyroscopic data to produce data from an animal's current state, but this has limitations. Take for example a normal standing posture. From an accelerometer's perspective, it closely resembles the "laying down" posture, but the posture can look very different if the animal is standing still, versus walking, versus running, and might look completely different from a "standing on incline" posture. A human trainer can instantly tell the difference between these postures and behaviors, but the process is much more difficult for a computer.

Training and handling working dogs is a costly process and requires specialized skills and techniques. Less subjective and lower-cost training techniques would not only improve our partnership with these dogs but also enable us to benefit from their skills more efficiently. To facilitate this, we are developing a canine body-area-network (cBAN) to combine sensing technologies and computational modeling to provide handlers with a more accurate interpretation for dog training. As the first step of this, we used inertial measurement units (IMU) to remotely detect the behavioral activity of canines. Decision tree classifiers and Hidden Markov Models were used to detect static postures (sitting, standing, lying down, standing on two legs and eating off the ground) and dynamic activities (walking, climbing stairs and walking down a ramp) based on the heuristic features of the accelerometer and gyroscope data provided by the wireless sensing system deployed on a canine vest. Data was collected from 6 Labrador Retrievers and a Kai Ken. The analysis of IMU location and orientation helped to achieve high classification accuracies for static and dynamic activity recognition.