CALDA: Improving Multi-Source Time Series Domain Adaptation With Contrastive Adversarial Learning.

Unsupervised domain adaptation (UDA) provides a strategy for improving machine learning performance in data-rich (target) domains where ground truth labels are inaccessible but can be found in related (source) domains. In cases where meta-domain information such as label distributions is available, weak supervision can further boost performance. We propose a novel framework, CALDA, to tackle these two problems. CALDA synergistically combines the principles of contrastive learning and adversarial learning to robustly support multi-source UDA (MS-UDA) for time series data. Similar to prior methods, CALDA utilizes adversarial learning to align source and target feature representations. Unlike prior approaches, CALDA additionally leverages cross-source label information across domains. CALDA pulls examples with the same label close to each other, while pushing apart examples with different labels, reshaping the space through contrastive learning. Unlike prior contrastive adaptation methods, CALDA requires neither data augmentation nor pseudo labeling, which may be more challenging for time series. We empirically validate our proposed approach. Based on results from human activity recognition, electromyography, and synthetic datasets, we find utilizing cross-source information improves performance over prior time series and contrastive methods. Weak supervision further improves performance, even in the presence of noise, allowing CALDA to offer generalizable strategies for MS-UDA.

A Survey of Unsupervised Deep Domain Adaptation.

Deep learning has produced state-of-the-art results for a variety of tasks. While such approaches for supervised learning have performed well, they assume that training and testing data are drawn from the same distribution, which may not always be the case. As a complement to this challenge, single-source unsupervised domain adaptation can handle situations where a network is trained on labeled data from a source domain and unlabeled data from a related but different target domain with the goal of performing well at test-time on the target domain. Many single-source and typically homogeneous unsupervised deep domain adaptation approaches have thus been developed, combining the powerful, hierarchical representations from deep learning with domain adaptation to reduce reliance on potentially-costly target data labels. This survey will compare these approaches by examining alternative methods, the unique and common elements, results, and theoretical insights. We follow this with a look at application areas and open research directions.

Robot-Enabled Support of Daily Activities in Smart Home Environments.

Smart environments offer valuable technologies for activity monitoring and health assessment. Here, we describe an integration of robots into smart environments to provide more interactive support of individuals with functional limitations. RAS, our Robot Activity Support system, partners smart environment sensing, object detection and mapping, and robot interaction to detect and assist with activity errors that may occur in everyday settings. We describe the components of the RAS system and demonstrate its use in a smart home testbed. To evaluate the usability of RAS, we also collected and analyzed feedback from participants who received assistance from RAS in a smart home setting as they performed routine activities.

Automated interpretation of influenza hemagglutination inhibition (HAI) assays: Is plate tilting necessary?

The hemagglutination inhibition assay (HAI) is widely used to evaluate vaccine-induced antibody responses as well as to antigenically characterize influenza viruses. The results of an HAI assay are based on an endpoint titration where the titers are generally manually interpreted and recorded by a well-trained expert. For serological applications, the lack of standardization in endpoint interpretation and interference from non-specific inhibitors in clinical samples can translate into a high degree of variability in the results. For example, tilting HAI plates at 45-60 degrees to look for a "tear drop pattern" with avian red blood cells is a common practice by many, but not all, research laboratories. In this work, we tested the hypothesis that an automated image analysis algorithm can be used to achieve an accurate and non-subjective interpretation of HAI assays-specifically without the need to tilt plates. In a side-by-side comparison study performed during FDA's biannual serological screening process for influenza viruses, titer calls for more than 2200 serum samples were made by the Cypher One automated hemagglutination analyzer without tilting and by an expert human with tilting. The comparison yielded 95.6% agreement between the expert reader and automated interpretation method (within ± 1 dilution) for the complete dataset. Performance was also evaluated relative to the type of red blood cell (turkey and guinea pig) and influenza strain (12 different viruses). For the subset that utilized guinea pig red blood cells (~44% of the samples), for which no plate tilting was required, the agreement with an expert reader was 97.2%. For the subset that utilized turkey red blood cells (~56% of the samples), for which plate tilting was necessary by the expert reader, the agreement was 94.3%. Overall these results support the postulate that algorithm-based interpretation of a digital record with no plate tilting could replace manual reading for greater consistency in HAI assays.