A5021435978- Neural dynamics and brain function
- Advanced Vision and Imaging
- Domain Adaptation and Few-Shot Learning
- Advanced Neural Network Applications
- Neural Networks and Applications
- Neurobiology and Insect Physiology Research
- Advanced Fluorescence Microscopy Techniques
- Visual perception and processing mechanisms
- Currency Recognition and Detection
- Brain Tumor Detection and Classification
- Accounting and Organizational Management
- Video Analysis and Summarization
- Color Science and Applications
- Identification and Quantification in Food
- Radiomics and Machine Learning in Medical Imaging
- Face Recognition and Perception
- Advanced Image and Video Retrieval Techniques
California Institute of Technology
2016-2024
Janelia Research Campus
2023-2024
Howard Hughes Medical Institute
2024
Green Valley High School
2023
Brown University
2016
Abstract We can now measure the connectivity of every neuron in a neural circuit, but we are still blind to other biological details, including dynamical characteristics each neuron. The degree which measurements alone inform understanding computation is an open question. Here show that with only network, predict activity underlying computation. constructed model network experimentally determined for 64 cell types motion pathways fruit fly optic lobe unknown parameters single and synapse...
We propose and systematically evaluate three strategies for training dynamically-routed artificial neural networks: graphs of learned transformations through which different input signals may take paths. Though some approaches have advantages over others, the resulting networks are often qualitatively similar. find that, in trained to classify images, layers branches become specialized process distinct categories images. Additionally, given a fixed computational budget, tend perform better...
Change is certain, but the what, where and when are unpredictable. E.g. approaching animals need to be detected promptly, also classified as friend or foe. What best way carry out detection discrimination jointly? How do humans it? We compared subjects' performance in three tasks: (1) with unpredictable stimulus onset time, (2) known (3) dual task: time. Four subjects participated a novel random-dot motion task, trial begins incoherent coherent appears after random delay. For different...
In 2020, over 250,000 people died from brain and Central Nervous System (CNS) tumors. Brain tumors account for around 85% or more of these. As 2021, 16 million in the United States have been diagnosed with some type cognitive impairment. The goal this paper is to find most effective set data augmentations correctly classify diseases deep networks, using structural Magnetic Resonance Imaging (MRI) data. This demonstrates a Greedy optimization technique sequence out blurring, distortion,...
We design an agent to search for frames of interest in video stored on a remote server, under bandwidth constraints. Using convolutional neural network score individual and hidden Markov model propagate predictions across frames, our accurately identifies temporal regions based sparse, strategically sampled frames. On subset the ImageNet-VID dataset, we demonstrate that using interpolate between frame scores allows requests 98% be omitted, without compromising frame-of-interest...
As I peruse the recent literature in OD and interact with colleagues who share my involvement theory practice am impressed by what perceive to be a movement toward increasing differentiation among many approaches field. This movement, large part, reflects maturation of With field has come view its constituent not as different ways doing same thing but rather things. The major issue no longer takes form debate defense effectiveness one approach against another search