Engineering Practice


Deep Builds

2025–2026 Continual Learning · Causal Inference

No-Exemplar Continual Learning via Causal Invariance

Neural networks forget old tasks when learning new ones. Instead of storing past examples, this asks why forgetting happens — and finds the answer in causal structure. Outperforms replay-based baselines with zero stored images.

PyTorch Continual Learning Causal Inference
2026 Causal Representation Learning · Recommendation

HMV-CRL: Separating Platform Influence from Genuine Preference

Recommendation engagement conflates what users want with what the algorithm pushed. This model learns both representations separately — and finds that platforms amplify engagement while suppressing the genuine preference signal.

PyTorch Causal Representation Learning Recommendation
2025–2026 Reinforcement Learning · Multi-Agent

Zero-Shot Coordination in Multi-Agent RL

Agents trained separately fail to coordinate when paired with a stranger — they built different habits. This asks what training structure gives agents conventions general enough to work with anyone.

PyTorch JAX PettingZoo
2025–2026 Causal Inference · Time Series

Causal State-Space Models for Time Series

Forecasting models predict what will happen — but decisions change the pattern. This integrates causal structure into sequence models so they can reason about interventions, not just correlations.

PyMC Mamba DoWhy
2025–2026 Physics-Informed ML · Dynamical Systems

Port-Hamiltonian Neural Networks

Neural networks predict physical motion but don't know energy is conserved — over long predictions, they drift. This bakes conservation laws into the architecture so violations are impossible.

torchdiffeq JAX Diffrax
2025 LLM Agents · Social Simulation

Generative Social Network Simulation

Social simulations use hand-coded rules. This uses language models as the agents themselves — each with distinct beliefs and tendencies — to study how opinions spread and polarize across a network.

GPT-4 NetworkX Mesa
2025 Vision-Language Models · Alignment

Personality-Aligned Vision-Language Model

People describe the same image differently depending on how they think. This aligns a vision-language model to specific personality styles, then tests whether that consistency makes it a better coaching assistant.

LLaVA TRL PEFT

Rapid Experiments

Loading…