Collective Graph Identification
Here, we solve several canonical problems in network analysis: entity resolution (determining when two observations correspond to the same entity), link prediction (inferring the existence of links), and node labeling (inferring hidden attributes).
Drug-Drug Interaction
Infer unknown drug drug interactions from a network of multiple drug-based similarities and known interactions.
https://github.com/linqs/psl-examples/tree/main/drug-drug-interaction
Synthetic Regression Generator
Used to help create and study ARIMA models. This laid a foundation for modeling noisy seasonal trends with moving averages.
GameAI
Work on heuristic search algorithms, expectimax pruning, reinforcment learning, and creating adversary bots.
Machine Learning
- Created infrastructures for many different ML pipelines.
- Also worked on minimizing branch speculation errors in the CPU; trained a perceptron model over different applications, used the outputs to make better predictions.
Ransomware Detection in the Hypervisor Layer (Nutanix Sponsored)
- Researched different ways of detecting ransomware for cheap and efficient use in the “cloud” layer.
- Trained a decision tree on a malware dataset.
- Applied the model in a hypervisor for Nutanix.