Algorithms for Sequential Decision Making

By Michael L. Littman. 1996

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Reinforcement Learning

Citations in the corpus

Bridging the Gap Between Value and Policy Based Reinforcement Learning (2017, methodology): “Previous work has used a Boltzmann weighted average operator [20, 5].”
Deep learning in neural networks: An overview (2014, background): “decision problems (POMDPs) (e.g., Schmidhuber, 1990d, 1991c; Ring, 1991, 1993, 1994; Williams, 1992a; Lin, 1993; Teller, 1994; Kaelbling et al., 1995; Littman et al., 1995; Boutilier and Poole, 1996; Littman, 1996; Jaakkola et al., 1995; McCallum, 1996; Kimura et al., 1997; Wiering and Schmidhuber, ”
Learning World Graphs to Accelerate Hierarchical Reinforcement Learning (2019, neutral): “At first glimpse, the world graph seems suitable for modelbased RL (Littman, 1996; Kaiser et al., 2019), but our method emphasizes the connections among neighboring pivotal states rather than transitions over any arbitrary pair, which is usually deemed as a much harder problem (Gu et al.”
Reinforcement learning (2019, neutral): “To evaluate the effect of the action taken, the agent also receives a reward which is a numeric evaluation of the quality of the action and its effect on the environment.”
Planning and Acting in Partially Observable Stochastic Domains (1998, methodology): “If the value functions are represented by sets of policy trees, the test for termination can be implemented exactly using linear programming [27].”
Planning treatment of ischemic heart disease with partially observable Markov decision processes (2000, background): “Unfortunately, although computable, the computational cost for doing this is enormous [22, 18, 20] and, in general, only POMDPs of low complexity can be solved exactly in practice.”
Adaptive computation and machine learning (1998, methodology): “The mathematical theory of PDS kernels in a general setting originated with the fundamental work of Mercer [1909] who also proved the equivalence of a condition similar to that of theorem 5.1 for continuous kernels with the PDS property. The connection between PDS and NDS kernels, in particular theo”
A Unified Analysis of Value-Function-Based Reinforcement-Learning Algorithms (1999, background): “Theorem 7 implies the convergence of Q-learning for alter nating rVlarkov games because min and max are both non-expansions (Littman 1996).”
A Survey of Learning in Multiagent Environments: Dealing with Non-Stationarity (2017, methodology): “The way to act optimally against the population was obtained by best responding using a cognitive hierarchy model (Camerer et al., 2004) which was modelled as a POMDP (Littman, 1996).”
On the undecidability of probabilistic planning and related stochastic optimization problems (2003, background): “The complexity of finite horizon MDPs and POMDPs has been extensively studied; see for example [26,38,40].”
Self-play Monte-Carlo tree search in computer poker (2014, background): “Its applicability to fully observable multi-agent Markov games has been explored by (Littman 1996).”
Strategy Iteration Is Strongly Polynomial for 2-Player Turn-Based Stochastic Games with a Constant Discount Factor (2010, background): “The .rst player tries to minimize the total expected discounted cost of the in.nite sequence of actions taken, with respect to a .xed discount factor.”
Symbiotic Evolution of Neural Networks in Sequential Decision Tasks (1997, background): “Several task properties deened by Littman (1996) are outlined below.”
Finding Approximate POMDP solutions Through Belief Compression (2011, unknown): “Littman showed a number of complexity bounds for POMDPs ( Littman, 1996 ); these bounds demonstrate the intractability of finding an exact solution for those POMDPs where a solution is known to exist.”