On-device research index

arXiv research

A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.

168,982 papers · 148 categories

Trend · papers per month

count % of corpus

3 results for “value-tracking”

Dynamics of Value-Tracking in Financial Marketsq-fin.TR

Value-tracking in financial markets breaks down when non-valuation-based traders dominate.

problem Understanding the threshold for value-tracking in financial markets.

method Simple discrete-time model to show how non-valuation-based traders can cause tracking errors.

result A threshold above which value-tracking breaks down without changes in asset value.

value-tracking financial markets tracking errors valuation-based traders non-valuation-based traders

2019-03-23abs ↗pdf ↗

Fast Value Tracking for Deep Reinforcement Learningstat.ML

New algorithm tracks deep RL value functions with uncertainty.

problem Static RL algorithms ignore uncertainty in agent-environment interactions.

method Langevinized Kalman Temporal-Difference (LKTD) algorithm using SGMCMC.

result LKTD algorithm converges to posterior distribution of deep RL parameters.

reinforcement learning deep learning Kalman filtering temporal-difference uncertainty quantification

2024-03-19abs ↗pdf ↗

Kalman meets Bellman: Improving Policy Evaluation through Value Trackingcs.LG

A new method improves policy evaluation in RL by tracking value uncertainties.

problem Limitations in existing policy evaluation methods for deep RL tasks.

method KOVA (Kalman Optimization for Value Approximation) based on extended Kalman filter.

result KOVA minimizes a regularized objective function that considers parameter and noisy return uncertainties.

Reinforcement Learning Policy Evaluation Kalman Filter Value Approximation Uncertainty

2020-02-17abs ↗pdf ↗