Search for: All records

Annotating automatic target recognition images is challenging; for example, sometimes there is labeled data in the source domain but no labeled data in the target domain. Therefore, it is essential to construct an optimal target domain classifier using the labeled information of the source domain images. For this purpose, we propose a transductive transfer learning (TTL) network consisting of an unpaired domain translation network, a pretrained source domain classifier, and a gradually constructed target domain classifier. We delve into the unpaired domain translation network, which simultaneously optimizes cycle consistency and modulated noise contrastive losses (MoNCE). Furthermore, the proposed hybrid CUT module integrated into the TTL network generates synthetic negative patches by noisy features mixup, and all the negative patches provide modulated weight into the NCE loss by considering similarity to the query. Apart from that, this hybrid CUT network considers query selection by entropy-based attention to specifying domain variants and invariant regions. The extensive analysis depicted that the proposed transductive network can successfully annotate civilian, military vehicles, and ship targets into the three benchmark ATR datasets. We further demonstrate the importance of each component of the TTL network through extensive ablation studies into the DSIAC dataset. 

We introduce caption-guided face recognition (CGFR) as a new framework to improve the performance of commercial-off-the-shelf (COTS) face recognition (FR) systems. In contrast to combining soft biometrics (e.g., facial marks, gender, and age) with face images, in this work, we use facial descriptions provided by face examiners as a piece of auxiliary information. However, due to the heterogeneity of the modalities, improving the performance by directly fusing the textual and facial features is very challenging, as both lie in different embedding spaces. In this paper, we propose a contextual feature aggregation module (CFAM) that addresses this issue by effectively exploiting the fine-grained word-region interaction and global image-caption association. Specifically, CFAM adopts a self-attention and a cross-attention scheme for improving the intra-modality and inter-modality relationship between the image and textual features, respectively. Additionally, we design a textual feature refinement module (TFRM) that refines the textual features of the pre-trained BERT encoder by updating the contextual embeddings. This module enhances the discriminative power of textual features with a cross-modal projection loss and realigns the word and caption embeddings with visual features by incorporating a visual-semantic alignment loss. We implemented the proposed CGFR framework on two face recognition models (ArcFace and AdaFace) and evaluated its performance on the Multi-Modal CelebA-HQ dataset. Our framework significantly improves the performance of ArcFace in both 1:1 verification and 1:N identification protocol. 

We propose an artificial iris to tackle sensitivity caused by photophobia. This artificial iris is made with a twisted nematic cell sandwiched between two linear polarizers. The light attenuation performance of a commercial TNC was compared with TNCs made for smart contact lenses. 

This survey paper provides an overview of the current state of Artificial Intelligence (AI) attacks and risks for AI security and privacy as artificial intelligence becomes more prevalent in various applications and services. The risks associated with AI attacks and security breaches are becoming increasingly apparent and cause many financial and social losses. This paper will categorize the different types of attacks on AI models, including adversarial attacks, model inversion attacks, poisoning attacks, data poisoning attacks, data extraction attacks, and membership inference attacks. The paper also emphasizes the importance of developing secure and robust AI models to ensure the privacy and security of sensitive data. Through a systematic literature review, this survey paper comprehensively analyzes the current state of AI attacks and risks for AI security and privacy and detection techniques. 

Paper and proposal deadlines are important milestones, conjuring up emotional memories to researchers. The question is if in the daily challenging world of scholarly research, deadlines truly incur higher sympathetic loading than the alternative. Here we report results from a longitudinal, in the wild study of n = 10 researchers working in the presence and absence of impeding deadlines. Unlike the retrospective, questionnaire-based studies of research deadlines in the past, our study is real-time and multimodal, including physiological, observational, and psychometric measurements. The results suggest that deadlines do not significantly add to the sympathetic loading of researchers. Irrespective of deadlines, the researchers’ sympathetic activation is strongly associated with the amount of reading and writing they do, the extent of smartphone use, and the frequency of physical breaks they take. The latter likely indicates a natural mechanism for regulating sympathetic overactivity in deskbound research, which can inform the design of future break interfaces. 

Abstract A steady supply of hosts at the susceptible stage for parasitism is a major component of mass rearing parasitoids for biological control programs. Here we describe the effects of storing 5th instar Plodia interpunctella larvae in dormancy on subsequent host development in the context of host colony maintenance and effects of the duration of host dormancy on the development of Habrobracon hebetor parasitoids reared from dormant hosts. We induced dormancy with a combination of short daylength (12L:12D) and lower temperature (15°C), conditions known to induce diapause in this species, and held 5th instar larvae of P. interpunctella for a series of dormancy durations ranging from 15 to 105 days. Extended storage of dormant 5th instar larvae had no significant impacts on survival, development, or reproductive potential of P. interpunctella , reinforcing that dormant hosts have a substantial shelf life. This ability to store hosts in dormancy for more than 3 months at a time without strong negative consequences reinforces the promise of using dormancy to maintain host colonies. The proportion of hosts parasitized by H. hebetor did not vary significantly between non-dormant host larvae and dormant host larvae stored for periods as long as 105 days. Concordant with a prior study, H. hebetor adult progeny production from dormant host larvae was higher than the number of progeny produced on non-dormant host larvae. There were no differences in size, sex ratio, or reproductive output of parasitoids reared on dormant hosts compared to non-dormant hosts stored for up to 105 days. Larval development times of H. hebetor were however longer when reared on dormant hosts compared to non-dormant hosts. Our results agree with other studies showing using dormant hosts can improve parasitoid mass rearing, and we show benefits for parasitoid rearing even after 3 months of host dormancy. 

Machining processes involve various sources of uncertainty which lead to inaccurate interpretation of results in the surface integrity of machined products. This work presents a physics-informed, data-driven modeling framework for achieving comprehensive uncertainty quantification (UQ) of the impact of process and material variability on machining-induced residual stress (RS). Uncertainty due to the variation in bulk material properties and model input parameters in machining are considered. Preliminary results showed that variations in calibration parameters have a substantial effect on modeling RS, while the variation in material properties has a smaller effect. Further research directions for UQ in machining are also outlined.