The Structure of Deception in Multi-Agent LLM Systems
How LLM agents lie, break promises, and exploit trust — a unified framework spanning taxonomy, empirical evaluation, and monitoring failure modes.
Why Multi-Agent Conversations Cannot Fix LLM Forecasting
Evidence from convergence analysis: conversational and mathematical calibration are functionally equivalent, converging to identical Brier scores.
Predictive Power of LLMs in Financial Markets
Does GPT beat BERT at extracting signal from the Fed's Beige Book? First place in the Meeting of the Minds Math Division.
Lamprey Population Dynamics with Adaptive Sex-Determination
A probabilistic differential-equation model of populations that adapt their sex ratio to food availability, predators, and habitat conditions.
Optimizing Movie Scheduling Solutions
Reframing theater film scheduling as a linear program that maximizes profitability — and explains why theaters rent movies instead of buying them.
Predicting Closing Prices — Optiver Kaggle Competition
Benchmarking deep learning against classical ML and linear regression for closing-price prediction across 200 stocks.
Sentiment Analysis and Stock Prices
SVM-based sentiment classification of market news at 77% accuracy — and why sentiment alone can't predict price moves.