arXiv subjects

Khoi D. Nguyen

Publications and source records attributed to Khoi D. Nguyen.

3 recordsLinked to original sources

arXiv2022

Inductive and Transductive Few-Shot Video Classification via Appearance and Temporal Alignments

We present a novel method for few-shot video classification, which performs appearance and temporal alignments. In particular, given a pair of query and support videos, we conduct appearance alignment via frame-level feature matching to achieve the appearance similarity score between the videos, while utilizing temporal order-preserving priors for obtaining the temporal similarity score between the videos. Moreover, we introduce a few-shot video classification framework that leverages the above appearance and temporal similarity scores across multiple steps, namely prototype-based training and testing as well as inductive and transductive prototype refinement. To the best of our knowledge, our work is the first to explore transductive few-shot video classification. Extensive experiments on both Kinetics and Something-Something V2 datasets show that both appearance and temporal alignments are crucial for datasets with temporal order sensitivity such as Something-Something V2. Our approach achieves similar or better results than previous methods on both datasets. Our code is available at https://github.com/VinAIResearch/fsvc-ata.

cs.CV↗

arXiv2022

POODLE: Improving Few-shot Learning via Penalizing Out-of-Distribution Samples

In this work, we propose to use out-of-distribution samples, i.e., unlabeled samples coming from outside the target classes, to improve few-shot learning. Specifically, we exploit the easily available out-of-distribution samples to drive the classifier to avoid irrelevant features by maximizing the distance from prototypes to out-of-distribution samples while minimizing that of in-distribution samples (i.e., support, query data). Our approach is simple to implement, agnostic to feature extractors, lightweight without any additional cost for pre-training, and applicable to both inductive and transductive settings. Extensive experiments on various standard benchmarks demonstrate that the proposed method consistently improves the performance of pretrained networks with different architectures.

cs.LG↗

arXiv2021

Rheological basis of skeletal muscle work loops

Skeletal muscle is subjected to simultaneous time-varying neural stimuli and length changes in vivo. Work loops are experimental representations of these in vivo conditions and exhibit force versus length responses that are not explainable using either soft matter rheology or the classical isometric and isotonic characterizations of muscle. These gaps in our understanding have often prompted the search for new muscle phenomena. However, we presently lack a framework to explain the mechanical origins of work loops that integrates multiple facets of current understanding of muscle, as a rheological material and also a stimulus-responsive actuator. Here we present a new hypothesis that work loops emerge by splicing together force versus length loops corresponding to different constant stimuli. Using published muscle datasets and a detailed sarcomere model, we find that the hypothesis accurately predicts work loops and helps understand them in terms of rheological behaviors measured at fixed-stimuli. Importantly, this framework identifies conditions under which a rheological understanding of muscle fails to explain the emergent work loops, and new muscle phenomena may be necessary to explain its in vivo function.

cond-mat.soft↗