The papers.
All first-author.
Research by YRI Fellows, listed the way a journal lists its contents. Published papers link straight to the index they appear in, so you can check each one yourself.
OR WATCH FELLOWS PRESENT THEIR RESEARCH ↓
Ruthwik D.··
Nidhi K.··
Aarnav B.··
Akul N.··
Devansh G.··
Aditya S.··
Raeyaan M.··
Aarav B.··
Sai T.··
Ishaan M.··
Suriya D.··
Ashwajit W.··
Sai P.··
Diya A.··
Atharv V.·
Srihaan E.··
Amaay G.··
Arjun V.··
Ayaan R.··
Neil V.··
Sidharth I.··
Viba M.··
Arfa H.··
Mohammed A.··
Aditya A.··
Krishik S.··
Naitik G.··
Khyati S.··
Svanika D.··
Devansh D.··
Elliot W.··
Karan A.··
Aaryan M.·
Arka S.·
Ishan S.·
Krishna K.·
Updated as results come in. PUBLISHED entries resolve in the publisher’s own index; ACCEPTED entries are scheduled for publication.
READ THE STUDENT STORIESThe research, in their own words
Every Fellow presents their work on camera at the end of the program. A few of the recent ones are below. Watch a fifteen-year-old walk through their own methodology: this is the part no brochure can fake.
Replay-Gated Cascade Consolidation: A Spiking-Network Model of Memory
Oral presentation at BICA 2026 in Merida, Mexico, recorded by the conference. A 1,000-neuron spiking network showing how replay drives memory consolidation, published in Springer Lecture Notes in Electrical Engineering.
Sai Srikar PatnaikCNN Grading of Knee X-Rays Against Expert Disagreement
Trained a compact CNN from scratch on 1,650 knee radiographs to grade osteoarthritis across the five Kellgren-Lawrence grades, finding that the model's largest confusion boundary aligned with the only boundary where human expert readers disagreed.
Cancer-Aware Attention CNN for Acute Lymphoblastic Leukemia Subtype Classification
Built an attention-based convolutional neural network for classifying acute lymphoblastic leukemia subtypes from blood smear images, using cancer-aware attention to focus the model on diagnostically relevant cell features.
AI-Powered Multi-Omics for Climate-Resilient Crops
Integrated transcriptomic, genomic and machine learning methods to identify stress-tolerant genetic targets in crops, an AI-powered multi-omics approach to breeding for a changing climate.
Ashneeth MandaOTFS vs OFDM for UAV Integrated Sensing and Communication
Compared OTFS and OFDM waveforms for integrated sensing and communication in UAV systems, evaluating which modulation scheme delivers better simultaneous radar sensing and data transmission in high-mobility drone environments.
AI Framework for Predicting Functional Roles of Non-Coding DNA
Built a multi-omics deep neural network that classifies regulatory elements in the non-coding 98 percent of the human genome as promoters or enhancers, then used SHAP explainable AI to open the black box, prioritizing 1,088 computationally identified enhancer candidates.
Endothelial ATAC-Seq in Cardiovascular and Neurodegenerative Disease
Applied ATAC-Seq to profile chromatin accessibility in endothelial cells, examining the regulatory changes that connect cardiovascular and neurodegenerative disease.
Matthew LiongCompiling WCA Megaminx Scrambles for a Two-Axis Machine
Developed a system for compiling official WCA Megaminx scrambles for a two-axis, four-primitive machine, bridging competitive cubing with computational algorithms.
AI-Driven EEG Depression Detection
Developed an AI-driven framework for detecting depression using EEG signal analysis, enabling non-invasive mental health screening.