The advancements in machine learning (ML) opened a new opportunity to bring intelligence
to the low-end Internet-of-Things (IoT) nodes, such as microcontrollers. Conventional ML …

Novel classes frequently arise in our dynamically changing world, eg, new users in the
authentication system, and a machine learning model should recognize new classes without …

Neural networks suffer from catastrophic forgetting when sequentially learning tasks phase-
by-phase, making them inapplicable in dynamically updated systems. Class-incremental …

Abstract Lifelong Person Re-identification (L-ReID) aims to learn from sequentially collected
data to match a person across different scenes. Once an L-ReID model is updated using …

From holistic low-dimension feature-based segmentation to deep polynomial neural
networks, Face Recognition (FR) accuracy has increased dramatically since its early days …

Asymmetric image retrieval, which typically uses small model for query side and large model
for database server, is an effective solution for resource-constrained scenarios. However …

In asymmetric retrieval systems, models with different capacities are deployed on platforms
with different computational and storage resources. Despite the great progress, existing …

The growing public concerns on data privacy in face recognition can be greatly addressed
by the federated learning (FL) paradigm. However, conventional FL methods perform poorly …

Rigorous and reproducible evaluation is critical for assessing the state of the art and for
guiding scientific advances in Artificial Intelligence. Evaluation is challenging in practice due …

Reducing inconsistencies in the behavior of different versions of an AI system can be as
important in practice as reducing its overall error. In image classification, sample-wise …