Protein structure prediction is a core challenge in computational biology, particularly for fragments within ligand-binding regions, where accurate modeling is still difficult. Quantum computing offers a novel first-principles modeling paradigm, but its application is currently limited by hardware constraints, high computational cost, and the lack of a standardized benchmarking dataset. In this work, we present QDockBank—the first large-scale protein fragment structure dataset generated entirely using utility-level quantum computers, specifically designed for protein–ligand docking tasks. QDockBank comprises 55 protein fragments extracted from ligand-binding pockets. The dataset was generated through tens of hours of execution on superconducting quantum processors, making it the first quantum-based protein structure dataset with a total computational cost exceeding one million USD. Experimental evaluations demonstrate that structures predicted by QDockBank outperform those predicted by AlphaFold2 and AlphaFold3 in terms of both RMSD and docking affinity scores. QDockBank serves as a new benchmark for evaluating quantum-based protein structure prediction.

Drug discovery is a challenging process, requiring the optimization of compounds to become safe and effective. Predicting molecular properties is an indispensable step in the drug discovery pipeline. Traditionally, this process is costly, involving multiple rounds of experiments, rendering it impractical for every candidate compound. Deep learning techniques have emerged as a promising approach to drug discovery to reduce the cost during the process. However, prevalent research in deep learning models focused on predicting molecular properties has primarily fixated on single-modal models, neglecting the potential benefits of combining different data modalities. To overcome this limitation, we introduce MRL-Mol: a deep Multimodal Representation Learning framework for accurate Molecular properties prediction. MRL-Mol harnesses three data modalities: sequence, graph, and image, augmenting the depth of comprehension. Leveraging a large-scale unlabeled dataset ( 1M unique molecules), we pretrain MRL-Mol to extract inter- and intra-modal information. Our study demonstrates the superior performance of MRL-Mol in predicting molecular properties across six benchmark datasets. Notably, MRL-Mol outperforms other state-of-the-art molecular properties prediction models. These findings suggest that by combining information from multiple data modalities, MRL-Mol can comprehend molecules better than single-modal deep learning models and identify molecular properties with better accuracy.

Deep learning techniques have emerged as valuable tools for video analysis and motion detection. Recent advancements in this field have shown promising results. Our objective is to leverage these video understanding techniques to aid teachers in evaluating their teaching quality and enhancing their effectiveness in the classroom. However, existing research on student behavior analysis primarily focuses on recognizing actions pertaining to classroom management, neglecting the identification of “learning behaviors” exhibited by students. To address this limitation, we introduce a novel video dataset specifically designed to capture the nuances of “learning behaviors” displayed by primary-grade students in the mathematics classroom, along with a dedicated student localization dataset focused on detecting the location of individuals. Our approach introduces a framework that utilizes deep learning-based object detection and action recognition techniques trained on our curated datasets to analyze and comprehend student learning behaviors in the classroom. To assess the performance of our approach, we conduct separate tests on our object detection and action recognition models. Sub-sequently, our framework is applied to a collection of recorded 360-degree classroom videos, enabling a thorough evaluation of its capabilities.

Unified systems for multi-sensor devices, particularly eye-tracking in Virtual Reality (VR), are intricate and often require the listening and streaming of multichannel data. In this project, we propose a visual analysis framework for replicating a participant's viewing involvement by interpreting head movements as rotations and point-of-gaze (POG) as on-screen indicators. Our solution suggests an additional layer of system for near-real-time for processing and analyzing this multi-device data to connect with the data and enable both near-real-time or subsequent offline viewing of the entire VR eye-tracking session. Moreover, our method provides a no-batteries-need solution to create traditional eye-tracking visualization techniques. Finally, we apply three prior education technology analysis metrics: higher density gaze for students, shorter fixation time, and less fixation duration variance for students to determine expertise levels in this system. We systematically establish a ubiquitous, multi-device, eye-tracking solution to incorporate this approach. We evaluate the effectiveness of our system through a user study, using both expertise and non- expertise levels, and selectively surveying to ascertain the quality of the replicated experience and we test the system by running a real-world user study with sixty four different participants. We demonstrate the application's significance and potential to integrate prior analysis metrics using the collected data which this data collection and analysis have been approved by IRB.

More and more hospitals are now deploying mul-tiple cameras in trauma room for a multi-perspective remote observation. but video surveillance system can cause privacy breach by showing and storing sensitive information of patients and staff. We use OpenPose which is a state-of-the-art human body skeletons estimation framework to extract 18 human key skeleton points. For privacy preservation, we can apply the image obfuscation techniques to human heads, we also can use human skeleton to replace the human body in the truth background. we proposed a head detection method based on the 5 key points of each head output from OpenPose. We applied the st-gcn algorithm to recognize human actions, we propose a interactive algorithm for multiple cameras to recognize and trace the same person in different cameras, Based on multi-view action recognition for the same person, we can take action recognition accuracy to a high level. Our experiment results prove that our proposed technique has a high performance in privacy protection applications. Now we focus on the interactive algorithm for multiple cameras.

Deep learning and data-driven approaches have shown great potential in scientific domains. The promise of data-driven techniques relies on the availability of a large volume of high-quality training datasets. Due to the high cost of obtaining data through expensive physical experiments, instruments, and simulations, data augmentation techniques for scientific applications have emerged as a new direction for obtaining scientific data recently. However, existing data augmentation techniques originating from computer vision yield physically unacceptable data samples that are not helpful for the domain problems that we are interested in. In this article, we develop new data augmentation techniques based on convolutional neural networks. Specifically, our generative models leverage different physics knowledge (such as governing equations, observable perception, and physics phenomena) to improve the quality of the synthetic data. To validate the effectiveness of our data augmentation techniques, we apply them to solve a subsurface seismic full-waveform inversion using simulated CO2leakage data. Our interest is to invert for subsurface velocity models associated with very small CO2leakage. We validate the performance of our methods using comprehensive numerical tests. Via comparison and analysis, we show that data-driven seismic imaging can be significantly enhanced by using our data augmentation techniques. Particularly, the imaging quality has been improved by 15% in test scenarios of general-sized leakage and 17% in small-sized leakage when using an augmented training set obtained with our techniques.