The long read: After his mother died, Otto became convinced he’d been wronged by the forensic investigators – and that one of them was a fraud. The truth was even stranger

Why are we still using PPO-like algorithms a decade later? A brief argument about generalization.

Multi-objective Reinforcement Learning (MORL) generalizes traditional reinforcement learning to scenarios involving multiple conflicting objectives that require explicit trade-off optimization. Thi...

e-petitions

Deep Q-networks use target networks to stabilise bootstrapped value learning, but the standard hard copy update also introduces a tradeoff. Holding the target network fixed, improves short term stabil...