Sound source localization (SSL) is the task of locating the source of sound within an image. Due to the lack of localization labels, the de facto standard in SSL has been to represent an image and audio as a single embedding vector each, and use them to learn SSL via contrastive learning. To this end, previous work samples one of local image features as the image embedding and aggregates all local audio features to obtain the audio embedding, which is far from optimal due to the presence of noise and background irrelevant to the actual target in the input. We present a novel SSL method that addresses this chronic issue by joint slot attention on image and audio. To be specific, two slots competitively attend image and audio features to decompose them into target and off-target representations, and only target representations of image and audio are used for contrastive learning. Also, we introduce cross-modal attention matching to further align local features of image and audio. Our method achieved the best in almost all settings on three public benchmarks for SSL, and substantially outperformed all the prior work in cross-modal retrieval.

Deep spiking neural networks (SNNs) have shown remarkable progress due to improvements, such as training algorithms. However, most of them have not considered the features of neuromorphic devices. Their improvement has relied on soft resets, which are computationally expensive and unsuitable for neuromorphic devices. To address this, this paper proposes gradient clipping in hard reset-based deep SNNs and explores how device noise enhances learning performance. According to our experiments on various datasets and models, the proposed approach improved the training performance of deep SNNs with hard reset. These findings bridge gaps between SNN algorithms and hardware constraints, paving the way for efficient neuromorphic computing.

This paper studies semi-supervised learning of video action detection (VAD), which assumes that only a small portion of training videos are labeled and the others remain unlabeled. The existing semi-supervised methods for VAD mainly focus on leveraging spatial context of unlabeled video, lacking its exploration of temporal context. To resolve this, we present a novel semi-supervised learning framework that effectively incorporates spatio-temporal context during training. We first introduce a new augmentation strategy called temporal cross-view augmentation to achieve robust representation across clips depicting the same action but not aligned on the time axis. We also propose a new context fusion method called global-local context fusion that effectively utilizes the spatio-temporal context of videos to enhances the features of each frame by incorporating those of other frames within a clip; this method aids in actively leveraging spatio-temporal context of video, leading to significant performance improvement. Our framework was evaluated on UCF101–24 and JHMDB-21, where it outperformed all existing methods in every evaluation setting.

Deepfakes, an increasingly prevalent form of information attack, pose serious threats to security and privacy. Deepfake voice attacks, in particular, have the potential to cause widespread disruption, creating an urgent need for an effective detection system. In this research, we propose the Trident of Poseidon - a novel set of triad training strategies aimed at enhancing the generalizability of deepfake voice detection models. Our solution comprises three key components: (1) Supervised Contrastive Learning, (2) Hard Negative Mining by Audio Re-synthesizing, and (3) Effective Proactive Batch Sampling. Together, these enable the model to learn more robust features. Our extensive experiments demonstrate that our approach outperforms existing methods in both in-domain and out-of-domain testing scenarios, making significant strides toward securing digital media against deepfake voice attacks.

To observe coastal waves in the surf zone, we develop a deep coastal wave tracking model using video imagery and apply it to Samcheok Beach on the east coast of Korea, which has a serious coastal erosion problem. A deep learning-based non-rigid image registration method is applied to detect spatial changes in waves at specific time intervals from a consecutive video image sequence and track their displacement. Through this displacement, the velocity of the propagated wave is estimated. For accurate detection of waves and tracking spatial changes, a preprocessing step is included in which only the sea area, excluding waves in the video imagery. The results of wave tracking are visually examined, using particles to see whether they follow the wave’s crest and white foam well, and it shows that the proposed method is qualitatively feasible for wave observation in a surf zone.

Recently, many large-scale neuromorphic systems that emulate spiking neural networks have been presented. Biological evidence emphasizes the importance of the log-normal distribution of biological neural and synaptic parameters in the brain; however, this fact is easily ignored sometimes, and the parameters are excessively optimized to scale up a system. This is because high-precision parameters require floating-point arithmetic–an operation known to consume high-energy and result in a high implementation cost in digital hardware. In this study, we propose a novel neuron implementation model that enhances neural and synaptic dynamics using the time-embedded floating-point arithmetic for better biological plausibility and low-power consumption. The proposed algorithm enables sharing temporal information with a membrane potential by time-embedded floating-point arithmetic, thus minimizing the memory usage of the neural state. In addition, this method need not access the static random-access memory at every time step, thus reducing the dynamic power consumption, even with a floating-point precision neural and synaptic dynamics. Using the proposed model, we implemented a core group with a total of 8,192 neurons on a field-programmable gate array device, Xilinx XC7K160T. The core group is designed for use in large-scale neuromorphic systems. We tested the neuron model in a core under various experimental conditions.

OBJECTIVE: To systematically review studies assessing the effects of health information technology (health IT) on patient safety outcomes. MATERIALS AND METHODS: The authors employed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement methods. MEDLINE, Cumulative Index to Nursing Allied Health (CINAHL), EMBASE, and Cochrane Library databases, from 2001 to June 2012, were searched. Descriptive and comparative studies were included that involved use of health IT in a clinical setting and measured effects on patient safety outcomes. RESULTS: Data on setting, subjects, information technology implemented, and type of patient safety outcomes were all abstracted. The quality of the studies was evaluated by 2 independent reviewers (scored from 0 to 10). A total of 69 studies met inclusion criteria. Quality scores ranged from 1 to 9. There were 25 (36%) studies that found benefit of health IT on direct patient safety outcomes for the primary outcome measured, 43 (62%) studies that either had non-significant or mixed findings, and 1 (1%) study for which health IT had a detrimental effect. Neither the quality of the studies nor the rate of randomized control trials performed changed over time. Most studies that demonstrated a positive benefit of health IT on direct patient safety outcomes were inpatient, single-center, and either cohort or observational trials studying clinical decision support or computerized provider order entry. DISCUSSION AND CONCLUSION: Many areas of health IT application remain understudied and the majority of studies have non-significant or mixed findings. Our study suggests that larger, higher quality studies need to be conducted, particularly in the long-term care and ambulatory care settings.

The TMO (Time-triggered Message-Triggered Object) structuring scheme is aimed at facilitating real-time (RT) distributed software engineering in a form which software engineers in the vast business software field can adapt to with small efforts. It is a syntactically simple and natural but semantically powerful extension of the conventional object structuring approaches. In the course of developing a TMO-structured RT system engineering methodology, an attractively simple approach to parallel and distributed RT simulation has been produced. Also, we recently implemented a TMO execution engine based on the Windows NT platform and named it TMOSM/NT. As an effort for validation of both the execution engine TMOSM/NT and the TMO-structured RT simulation approach, a freeway automobile traffic simulator named distributed object-oriented freeway simulator (DOFS), has been constructed. In this paper, the design techniques and tools applied in development of DOFS are discussed.