A5074077264- Ultrasound Imaging and Elastography
- Photoacoustic and Ultrasonic Imaging
- Ultrasound and Hyperthermia Applications
- Ultrasonics and Acoustic Wave Propagation
- Coronary Interventions and Diagnostics
- Protein Structure and Dynamics
- Advanced MRI Techniques and Applications
- Tissue Engineering and Regenerative Medicine
- Cardiovascular Health and Disease Prevention
- Cardiac Imaging and Diagnostics
- Machine Learning in Materials Science
- Voice and Speech Disorders
- Generative Adversarial Networks and Image Synthesis
- Bioinformatics and Genomic Networks
- Congenital Heart Disease Studies
- 3D Printing in Biomedical Research
- Big Data and Business Intelligence
- Computational Drug Discovery Methods
- Additive Manufacturing and 3D Printing Technologies
- Innovative Microfluidic and Catalytic Techniques Innovation
- Artificial Intelligence in Healthcare and Education
- Educational Robotics and Engineering
- Mechanical Circulatory Support Devices
- Ultrasound and Cavitation Phenomena
- Advanced Neuroimaging Techniques and Applications
Massachusetts Institute of Technology
2022-2024
Georgia Institute of Technology
2019-2023
The Wallace H. Coulter Department of Biomedical Engineering
2019-2023
Moscow Institute of Thermal Technology
2022
Henan University of Science and Technology
2022
Emory University
2020-2021
Xi'an Jiaotong University
2014-2018
Ministry of Education of the People's Republic of China
2017
Predicting the binding structure of a small molecule ligand to protein -- task known as molecular docking is critical drug design. Recent deep learning methods that treat regression problem have decreased runtime compared traditional search-based but yet offer substantial improvements in accuracy. We instead frame generative modeling and develop DiffDock, diffusion model over non-Euclidean manifold poses. To do so, we map this product space degrees freedom (translational, rotational,...
The biological functions of proteins often depend on dynamic structural ensembles. In this work, we develop a flow-based generative modeling approach for learning and sampling the conformational landscapes proteins. We repurpose highly accurate single-state predictors such as AlphaFold ESMFold fine-tune them under custom flow matching framework to obtain sequence-conditoned models protein structure called AlphaFlow ESMFlow. When trained evaluated PDB, our method provides superior combination...
Molecular conformer generation is a fundamental task in computational chemistry. Several machine learning approaches have been developed, but none outperformed state-of-the-art cheminformatics methods. We propose torsional diffusion, novel diffusion framework that operates on the space of torsion angles via process hypertorus and an extrinsic-to-intrinsic score model. On standard benchmark drug-like molecules, generates superior ensembles compared to methods terms both RMSD chemical...
Abstract Generative AI is rapidly transforming the frontier of research in computational structural biology. Indeed, recent successes have substantially advanced protein design and drug discovery. One key methodologies underlying these advances diffusion models (DM). Diffusion originated computer vision, taking over image generation offering superior quality performance. These were subsequently extended modified for uses other areas including DMs are well equipped to model high dimensional,...
Protein structure prediction has reached revolutionary levels of accuracy on single structures, yet distributional modeling paradigms are needed to capture the conformational ensembles and flexibility that underlie biological function. Towards this goal, we develop EigenFold, a diffusion generative framework for sampling distribution structures from given protein sequence. We define process models as system harmonic oscillators which naturally induces cascading-resolution along eigenmodes...
This book begins with a detailed introduction to the fundamental principles and historical development of GANs, contrasting them traditional generative models elucidating core adversarial mechanisms through illustrative Python examples. The text systematically addresses mathematical theoretical underpinnings including probability theory, statistics, game theory providing solid framework for understanding objectives, loss functions, optimisation challenges inherent GAN training. Subsequent...
Laser-activated bioprobes with high photothermal conversion efficiency (IRPDA@PFH NDs) based on biocompatible IR-780 doped polydopamine perfluorocarbon nanodroplets (NDs) were developed. When protected by gelatin microspheres, their near-spherical morphologies can be easily observed transmission electron microscope. Doping (3 w/w % of added dopamine hydrochloride) significantly enhance near-infrared (NIR) absorption and to 57.7%. The enhanced NIR nonradiative relaxation are preferred...
Abstract The heart is the first organ to develop in human embryo through a series of complex chronological processes, many which critically rely on interplay between cells and dynamic microenvironment. Tight spatiotemporal regulation these interactions key development diseases. Due suboptimal experimental models, however, little known about role microenvironmental cues development. This study investigates use 3D bioprinting perfusion bioreactor technologies create bioartificial constructs...
The recent developments of foundation models in computer vision, especially the Segment Anything Model (SAM), allow scalable and domain-agnostic image segmentation to serve as a general-purpose tool. In parallel, field medical has benefited significantly from specialized neural networks like nnUNet, which is trained on domain-specific datasets can automatically configure network tailor specific challenges. To combine advantages models, we present nnSAM, synergistically integrates SAM model...
Vascular atresia are often treated via transcatheter recanalization or surgical vascular anastomosis due to congenital malformations coronary occlusions. The cellular response is, however, largely unknown and current techniques rely on restoration rather than optimization of flow into the atretic arteries. An improved understanding post may result in reduced restenosis. Here, an vitro platform is used model pulmonary arteries (PAs) for procedural planning reduce Bifurcated PAs bioprinted...
Transcranial ultrasound imaging and therapy depend on the efficient transmission of acoustic energy through skull. Multiple previous studies have concluded that a large incidence angle should be avoided during transcranial-focused to ensure Alternatively, some other shown longitudinal-to-shear wave mode conversion might improve skull when is increased above critical (i.e., 25° 30°).The effect porosity at varying angles was investigated for first time elucidate why decreased in cases but...
The axial resolution of ultrasonic imaging is confined by the temporal width acoustic pulse generated transducer, which has a limited bandwidth. Deconvolution can eliminate this effect and, therefore, improve resolution. However, most methods perform deconvolution scan line line, and therefore information embedded within neighbor lines unexplored, especially for those materials with layered structures such as blood vessels. In paper, joint sparse representation model proposed to increase...
Computational methods that operate on three-dimensional molecular structure have the potential to solve important questions in biology and chemistry. In particular, deep neural networks gained significant attention, but their widespread adoption biomolecular domain has been limited by a lack of either systematic performance benchmarks or unified toolkit for interacting with data. To address this, we present ATOM3D, collection both novel existing benchmark datasets spanning several key...
Artificial Intelligence (AI) has permeated numerous aspects of our daily lives, from predictive text on smartphones to complex decision-making systems in healthcare and finance. While AI shown remarkable accuracy efficiency, it is often criticized for being a 'black box,' particularly when comes models like deep learning large language (LLMs). This where Explainable (XAI) into play.Explainable aims make decisions transparent, understandable, interpretable. The lack interpretability raised...
For the purpose of noninvasively visualizing dynamics contact between vibrating vocal fold medial surfaces, an ultrasonic imaging method which is referred to as array-based transmission glottography proposed. An array ultrasound transducers used detect wave transmitted from one side folds other through small-sized folds. A passive acoustic mapping employed visualize and locate contact. The results investigation using tissue-mimicking phantoms indicate that it feasible use proposed soft...
Pulse-inversion subharmonic (PISH) imaging can display information relating to pure cavitation bubbles while excluding that of tissue. Although plane-wave-based ultrafast active (UACI) monitor the transient activities bubbles, its resolution and cavitation-to-tissue ratio (CTR) are barely satisfactory but be significantly improved by introducing eigenspace-based (ESB) adaptive beamforming. PISH UACI a natural combination for activity in tissue; however, it raises two problems: 1) ESB...