New method infers intent from suboptimal behavior by modeling incorrect internal beliefs about dynamics.
problem Inferring intent from suboptimal human behavior using traditional methods assumes near-optimality, which is not always valid.
method Model suboptimal behavior as internal model misspecification, estimating incorrect beliefs about dynamics.
result Accurately models human intent by accounting for internal model inaccuracies.
Novel IRL method identifies suboptimal medical decisions in ICU data.
problem Identifying suboptimal medical decisions in clinical settings.
method Incorporates Inverse Reinforcement Learning with a pruning step to identify and remove suboptimal actions.
result Pruning step effectively identifies clinical priorities and values from suboptimal data.
In many platforms, user arrivals exhibit a self-reinforcing behavior: future user arrivals are likely to have preferences similar to users who were satisfied in the past. In other words, arrivals exhibit positive externalities. We study multiarmed bandit (MAB) problems with positive externalities. We show that the self…
Paper explores limits of imitation learning in MDPs, setting new suboptimality bounds.
problem Understanding the statistical limits of imitation learning in MDPs.
method Analyzes minimax statistical limits in two settings: pre-interaction and interaction.
result Establishes suboptimality bounds for imitation learning in MDPs, showing improvements with knowledge of transition.
Paper develops a framework for learning interpretable representations of sequential decision behavior.
problem Obtaining a transparent description of existing behavior.
method Inverse decision modeling framework, formalizing both forward and inverse problems.
result Learning interpretable representations of behavior, including suboptimal actions, biased beliefs, and imperfect knowledge.
New algorithm tackles confounding in offline RL for partially observable MDPs.
problem Confounding in offline reinforcement learning for partially observable MDPs.
method P3O algorithm using proximal causal inference and pessimistic confidence regions.
result Achieves n−1/2-suboptimality for offline RL in POMDPs with confounded data. KL-regularized RL from expert demos can lead to slow, unstable learning.
problem Pathological training dynamics in KL-regularized RL from expert demonstrations.
method Empirical analysis and non-parametric behavioral reference policies.
result KL-regularized RL can be significantly improved by using non-parametric behavioral policies.
Two entropy measures quantify suboptimal portfolio performance.
problem Measuring suboptimality in investment portfolios.
method Relative entropy (KL divergence) calculations.
result Suboptimal portfolios appear better than Kelly portfolios under certain measures.
Active Inverse Reward Design improves AI agent training by querying users for reward function preferences.
problem Iterative reward function tuning in AI agents is inefficient and may not generalize well.
method Structured queries to the user to compare reward functions, updating posterior with IRD.
result Substantially outperforms IRD in test environments, inferring non-linear rewards.
New algorithm learns policies without uniform overlap assumption.
problem Learning optimal policies from non-uniformly collected data.
method Pessimistic Policy Learning (PPL) using lower confidence bounds.
result Efficient policy learning for adaptively collected data.
Boltzmann exploration is a classic strategy for sequential decision-making under uncertainty, and is one of the most standard tools in Reinforcement Learning (RL). Despite its widespread use, there is virtually no theoretical understanding about the limitations or the actual benefits of this exploration scheme. Does it…
Improves SGD convergence by online linear regression of gradients in multiple directions.
problem SGD's rough approximations of gradients lead to suboptimal plateaus and saddle points.
method Online linear regression of noisy gradients using PCA to estimate second-order behavior, and gradient descent outside the subspace.
result Improves convergence by avoiding suboptimal plateaus and saddles, leading to better final values.
Interpole learns transparent decision-making policies from data.
problem Understanding human decision-making in opaque environments.
method Interpole combines belief-update and belief-action mapping estimation.
result Interpole provides interpretable models of decision-making behavior.
New suboptimal algorithm for best subset selection in high-dimensional data.
problem Nonconvex and computationally challenging best subset selection in linear regression.
method Introducing a new suboptimal algorithm and comparing it with other popular methods.
result The new procedure is a competitive suboptimal algorithm for high-dimensional data.
Paper tackles RLHF with DCPPO method, proving near-optimal suboptimality.
problem Challenges in offline RLHF with limited human feedback and bounded rationality.
method DCPPO method involving three stages: MLE, reward function recovery, and pessimistic value iteration.
result DCPPO's suboptimality almost matches classical pessimistic offline RL in terms of distribution shift and dimension.
The maximum a posteriori (MAP) configuration of binary variable models with submodular graph-structured energy functions can be found efficiently and exactly by graph cuts. Max-product belief propagation (MP) has been shown to be suboptimal on this class of energy functions by a canonical counterexample where MP conver…
The notion that economies should normally be in equilibrium is by now well-established; equally well-established is that economies are almost never precisely in equilibrium. Using a very general formulation, we show that under dynamics that are second-order in time a price system can remain away from equilibrium with p…
New stability analysis improves generalization of multipass SGD.
problem Improper preconditioning affects generalization in multipass SGD.
method Developed on-average stability analysis for multipass SGD.
result Proper preconditioning yields optimal effective dimension dependence.
Improved adaptive rates for Lipschitz bandit problem.
problem Sequentially maximize an unknown Lipschitz function with noisy evaluations.
method Characterizes regret through integrals of suboptimality gaps over level sets, adapting to local growth.
result Improved adaptive rates strictly improving over classical zooming bounds.
The Lasso is suboptimal in sparse linear regression due to design matrix constraints.
problem The suboptimality of the Lasso estimator in sparse linear regression.
method Characterization of optimal estimation rates and analysis of the Lasso estimator.
result The Lasso is provably minimax rate-suboptimal when the minimum singular value is small.
New algorithm for MIPS without preprocessing, controlling suboptimality.
problem Fast query time and suboptimality control for MIPS.
method Casting MIPS as a Best Arm Identification problem, introducing a new bandit setting.
result Outperforms state-of-the-art methods on synthetic and real-world datasets.
New bounds on adaptivity cost in stochastic optimization.
problem Understanding the cost of changing strategies in stochastic optimization.
method Proving impossibility results for adaptivity in non-smooth stochastic convex optimization.
result Lower bounds on the price of adaptivity for different levels of uncertainty.
Revisits VIC method to correct intrinsic reward bias in stochastic environments.
problem Intrinsic reward bias in VIC leading to suboptimal solutions.
method Proposes two methods based on transitional probability model and Gaussian mixture model to correct bias.
result Achieves maximal empowerment through corrected intrinsic reward.
Proposes a new algorithm to improve Frank-Wolfe method efficiency.
problem Improves Frank-Wolfe method's stability and efficiency.
method Introduces 1-SFW, a one-sample stochastic Frank-Wolfe algorithm.
result Achieves optimal convergence rate and first-order stationary point.
We study a novel machine learning (ML) problem setting of sequentially allocating small subsets of training data amongst a large set of classifiers. The goal is to select a classifier that will give near-optimal accuracy when trained on all data, while also minimizing the cost of misallocated samples. This is motivated…
PE-SVI reduces SVI inference complexity by finding a suitable start point.
problem Complex posterior inference in graphical models leads to suboptimal learning.
method PE-SVI uses a pseudo-encoded start point to reduce gradient steps and step sizes.
result PE-SVI achieves the same ELBo objective as SVI with less than 1% of the required steps.
Practical model building processes are often time-consuming because many different models must be trained and validated. In this paper, we introduce a novel algorithm that can be used for computing the lower and the upper bounds of model validation errors without actually training the model itself. A key idea behind ou…
Enhances reinforcement learning from sparse data.
problem Limited data for offline reinforcement learning.
method Trajectory-based data augmentation.
result Improves reinforcement learning performance.
New algorithm reduces suboptimality in imitation learning to nearly optimal levels.
problem Statistical limits of imitation learning in MDPs with known transitions.
method Mimic-MD algorithm and reduction to value estimation problem.
result Upper bound of O(∣S∣H3/2/N) for suboptimality, with efficient computation. Study on decision-making cascades with agents having varying beliefs and noise levels.
problem Optimizing decision-making in a cascade of agents with heterogeneous beliefs and noise.
method Recursive belief update and analysis of optimal decision rules, predecessor selection problem characterization.
result Optimal decisions can deviate from true prior beliefs in certain conditions, highlighting the importance of social learning.
Method infers internal model from rational foraging behavior.
problem Discover how the brain represents and computes dynamic beliefs.
method POMDP model, EM algorithm, maximum likelihood estimation.
result Successfully recovered internal model and reward function of foraging agents.
Optimizes decisions under strategic individual behavior.
problem Optimal decision-making in strategic environments.
method Characterizes strategic effort, identifies optimal policies under monotonic cost assumptions, develops iterative search algorithm.
result Demonstrates higher utility of decision policies accounting for strategic behavior.
A new forecasting framework uses suboptimal embeddings to improve multivariate time series prediction.
problem Randomly selecting embeddings or brute force methods often lead to suboptimal forecasts.
method Develops a framework that uses various suboptimal embeddings obtained via combinatorial optimization.
result Achieves the best results among existing frameworks for various datasets.
New method improves combinatorial optimization by overcoming inefficient sampling.
problem Wandering in contours: sampling similar solutions for long periods.
method Reheated Gradient-based Discrete Sampling with a novel mechanism.
result Superior performance across various combinatorial optimization problems.
New study shows exponential lower bound for RL even with constant suboptimality gap.
problem Can RL be sample-efficient with a constant suboptimality gap?
method Analyzes reinforcement learning in the online setting with a linearly realizable optimal Q-function.
result An exponential sample complexity lower bound still holds even with a constant suboptimality gap.
Least Squares Estimators are suboptimal for 5D convex functions.
problem Suboptimality of Least Squares Estimators in estimating multidimensional convex functions.
method Analysis of natural subclasses of convex functions in random and fixed design settings.
result Risk of LSE is n−2/d while minimax risk is n−4/(d+4) for d≥5. Training a source model optimally for its own task is suboptimal for downstream transfer.
problem The optimality of a source model for its own task hinders downstream transfer performance.
method Analyzes L2-SP ridge regression, characterizes transfer-optimal source penalty, and identifies alignment-dependent effects.
result Transfer benefits from stronger source regularization when aligned imperfectly, and from weaker regularization when aligned perfectly.
AC-Teach uses an ensemble of suboptimal teachers to improve exploration in RL.
problem Improving exploration efficiency in long-horizon tasks with sparse rewards.
method Bayesian Actor-Critic with an ensemble of suboptimal teachers.
result AC-Teach improves sample efficiency over baselines on various tasks.
A new measure helps compute suboptimality in entropy-regularized methods.
problem Computing suboptimality in entropy-regularized variational objectives when unnormalised densities are unavailable.
method Introduced 'kernel gradient discrepancy' (KGD) to compute suboptimality explicitly.
result KGD characterizes kernel Stein discrepancy (KSD) in the standard Bayesian context and measures variational gradient size.
Algorithm improves learning by integrating diverse agents' behaviors.
problem Lack of social learning in reinforcement learning algorithms.
method Free energy approach for social bandit learning.
result Algorithm converges to optimal policy and enhances learning.
New job recommendation system improves job seekers' welfare through field experiments.
problem Current job recommendation systems focus on clicks and applications, not job seekers' welfare.
method Developed a job-search model with two dimensions: utility and success probability. Conducted field experiments to validate model predictions.
result Welfare-optimal job recommendation algorithms outperform existing approaches and perform close to the benchmark.
Diffusion-QL uses diffusion models to improve offline RL performance.
problem Offline RL struggles with function approximation errors on out-of-distribution actions.
method Diffusion-QL represents the policy as a conditional diffusion model and optimizes action-values.
result Diffusion-QL achieves state-of-the-art performance on D4RL benchmark tasks.
Spectral algorithms improve under covariate shift with novel weighted techniques.
problem Improving spectral algorithms' performance under covariate shift.
method Analysis of spectral algorithms in non-parametric regression over RKHS, proposing a weighted spectral algorithm with clipped weights.
result Normalized weighted spectral algorithm achieves optimal capacity-independent convergence rates, and clipped weights can approach optimal capacity-dependent rates.
Wide neural networks have non-attracting local minima.
problem Understanding the impact of suboptimal local minima in deep and wide neural networks.
method Construction of non-attracting local minima and saddle points in wide neural networks.
result Wide neural networks have non-attracting local minima, even though they are not negatively impacted by suboptimal local optima.
Proposes a new query autocompletion method that maximizes retrieval performance.
problem Users often select suboptimal queries due to unknown best retrieval performance.
method Formulates query autocompletion as ranking item rankings, uses counterfactual learning.
result Empirical results show improved query suggestions for better retrieval performance.
New method identifies flawed internal models of the world in animals.
problem How animals make decisions with partial sensory information.
method Generalizes Inverse Rational Control to continuous nonlinear dynamics and noise.
result Identifies the best internal model explaining an agent's actions.
IDT learns human preferences from uncertain decisions, even when humans are suboptimal.
problem Learning human preferences from uncertain and suboptimal decisions.
method Inverse decision theory (IDT) framework, statistical analysis of IDT, characterizing sample complexity.
result Learning preferences is easier when decisions are more uncertain, even if humans are suboptimal.
Paper explores PG for MCR, finding suboptimal policies but providing bounds.
problem Optimizing coherent risk in MDPs using PG.
method Developed practical PG implementation using state distribution reweighting.
result PG can learn risk-sensitive policies but suboptimality is common.