αlphabell.
α · People · Anders Lindholm
Stylised avatar of Anders Lindholm

Anders Lindholm

Long-tenured contributor

Based in Helsinki
Node-cell riemann-44
ORCID 0000-2335-7300-4771

Research

Anders leads riemann-44, the cell working on time-correlated reward learning and long-horizon credit assignment in world-model-trained policies. The cell sits at the boundary between the world-models axis and the agentic-engineering axis on the question of how a learned dynamics model interacts with reward signals that arrive on time scales much longer than typical RL training horizons.

He has been with alphabell since 2021 and is a co-author on the counterfactual-rollouts paper. The riemann-44 line of work is one of the more theory-leaning threads in the world-models axis.

Anders is based in Helsinki. He is one of the contributors who works closely with the Nordic regional cluster of universities.

Background

Ph.D. machine learning, Aalto University, 2015.

Prior to alphabell: Aalto; Volterra Cognition; Wayfarer Institute.

Selected publications

Full publications index →

Recent talks

  • Long-horizon credit assignment in world models, NeurIPS 2025
  • Time-correlated rewards and dynamics, RLDM 2024
Working with

Anders is currently part of node-cell riemann-44, working under the World models research axis. The cell is open to substantive correspondence from researchers working on adjacent problems; route requests through riemann-44@alphabell.com or directly to Anders at anders-lindholm@alphabell.com.