Active learning framework for optimizing human preferences in reinforcement learning.
problem Selecting most informative feedback for training models of human preferences.
method Proposes an active learning framework to collect preferential feedback online or offline.
result Errors in DPO logit estimates diminish with more feedback.
Advocates a local feedback approach for RL in unknown systems.
problem Finding optimal feedback laws in unknown nonlinear dynamical systems.
method Searches over a local feedback representation consisting of an open-loop sequence and an optimal linear feedback law.
result Results in highly efficient training and superior performance compared to global methods.
Proposes a TS approach for Bayesian optimization with preferential feedback.
problem Optimizing with preference feedback in complex applications.
method Uses Thompson Sampling with a dueling kernel and anchor invariance.
result Performance matches standard TS for scalar feedback in finite time.
Learning optimal feedback control laws capable of executing optimal trajectories is essential for many robotic applications. Such policies can be learned using reinforcement learning or planned using optimal control. While reinforcement learning is sample inefficient, optimal control only plans an optimal trajectory fr…
Study evaluates new models using human feedback from another model.
problem Evaluate a new model using human feedback collected for another model.
method Formalize problem, propose model-based and model-free estimators, analyze unbiasedness, and empirically evaluate.
result Proposed estimators can predict absolute values, rank, and optimize evaluated policies.
New RL algorithm optimizes policies with bandit feedback, matching previous bounds.
problem Optimizing policies with unknown transitions and bandit feedback.
method Optimistic Trust Region Policy Optimization (TRPO) algorithm.
result Sub-linear regret bounds for both stochastic and adversarial rewards.
Unified framework for analyzing online convex optimization across various settings.
problem Analyzing online convex optimization in different settings and feedback types.
method Unified framework allowing systematic proposal and analysis of meta-algorithms.
result Comparable regret bounds for various feedback types and adversary types.
A new one-point feedback scheme improves ZO algorithms for black-box optimization.
problem Optimizing black-box functions without gradient information.
method Proposes a one-point feedback scheme to estimate gradients using residuals.
result Matches query complexity of two-point schemes for deterministic Lipschitz functions.
Paper studies CLO with partial feedback, improving decision-making in uncertain contexts.
problem Improving decision-making in contexts with uncertain cost coefficients using partial feedback.
method Unified class of offline learning algorithms for CLO with different types of feedback, using IERM framework.
result Fast-rate regret bound for IERM with partial feedback and misspecified model classes.
Study on sample complexity for pure exploration in feedback graph settings.
problem Sample complexity of pure exploration in online learning with feedback graphs.
method Derive instance-specific lower bounds and present asymptotically optimal algorithm TaS-FG.
result TaS-FG is asymptotically optimal and efficient across different graph configurations.
PILAF optimizes reward models from human feedback for better policy alignment.
problem Creating accurate reward models from human feedback for policy optimization.
method Policy-Interpolated Learning for Aligned Feedback (PILAF) that explicitly aligns preference learning with maximizing underlying oracle reward.
result PILAF is optimal from both optimization and statistical perspectives, demonstrating strong performance in RLHF settings.
PCTS optimizes noisy, delayed, multi-fidelity feedbacks in black-box optimization.
problem Optimizing unknown functions with noisy, delayed, and multi-fidelity feedbacks.
method ProCrastinated Tree Search (PCTS) with DUCB1 and DUCBV algorithms.
result PCTS achieves better regret bounds for delayed, noisy, and multi-fidelity feedbacks.
Efficient boosting method for regression with limited feedback.
problem Online boosting for regression tasks with noisy multi-point bandit feedback.
method Efficient regret minimization method with online boosting algorithm and projection-free online convex optimization.
result Improved state-of-the-art guarantees in efficiency.
Study optimal arms in combinatorial bandits with semi-bandit feedback and finite budget.
problem Finding optimal arms in combinatorial bandits with semi-bandit feedback and finite budget constraints.
method Proposes a generic algorithm covering various arm elimination strategies and derives lower bounds.
result Demonstrates sufficient and necessary budget requirements for finding the best arm.
New method learns from either positive or negative feedback alone.
problem Limited applicability of existing preference optimization methods in scenarios with only unpaired feedback.
method Decouples learning from positive and negative feedback, using expectation-maximization (EM) to optimize probability of positive outcomes and explicitly incorporate negative examples.
result Stable learning from negative feedback alone demonstrated.
The paper addresses learner privacy in convex optimization with feedback.
problem Privacy risks from eavesdropping adversaries observing learner's queries.
method Optimally obfuscating learner's queries to make their learned optimal value hard to estimate.
result Query complexity overhead is additive in L in the minimax formulation, multiplicative in L in the Bayesian formulation. This paper tackles combinatorial optimization under uncertainty with limited feedback.
problem Tackling combinatorial optimization problems with uncertain or unknown parameters.
method Review of techniques for combinatorial pure exploration with limited bandit feedback.
result Introduction of methods for combinatorial optimization under uncertainty with limited observation.
New framework PBBO optimizes latent functions with preferential feedback.
problem Optimizing latent functions with parallel preferential feedback.
method Gaussian process model with a likelihood for parallel comparisons.
result Framework PBBO generalizes and expands Bayesian optimization.
SLHF uses sequential game theory to optimize preferences from human feedback.
problem Optimizing preferences from human feedback in sequential settings.
method SLHF frames the problem as a sequential-move game between Leader and Follower, decomposing the optimization into refinement and adversarial optimization.
result SLHF achieves strong alignment across diverse preference datasets and scales to large models.
Paper tackles combinatorial reinforcement learning with preference feedback.
problem Modeling long-term user engagement in scenarios like recommender systems and online advertising.
method Assumes a contextual MNL preference model with linear mean utilities and approximates item values. Proposes MNL-VQL algorithm.
result Achieves nearly minimax-optimal regret for linear MDPs with preference feedback.
QS-BO optimizes functions using only rank-based feedback.
problem Optimizing expensive functions with unreliable or unavailable metric values.
method Quantile-scaling pipeline to convert ranks into Gaussian targets.
result QS-BO consistently achieves lower objective values and is statistically significant.
Paper introduces a new framework for optimizing non-convex functions.
problem Optimizing non-convex functions, especially DR-submodular and concave functions.
method Developed a general meta-algorithm to convert linear/quadratic optimization to optimization of upper-linearizable/quadratizable functions.
result Unified approach to concave and DR-submodular optimization problems.
DUET optimizes LLM training data mixtures using feedback from unseen tasks.
problem Optimizing LLM performance on unseen tasks without prior data knowledge.
method Interleaves influence function with Bayesian optimization for data selection.
result DUET converges to optimal training data mixture for unseen tasks.
This work focuses on dynamic regret of online convex optimization that compares the performance of online learning to a clairvoyant who knows the sequence of loss functions in advance and hence selects the minimizer of the loss function at each step. By assuming that the clairvoyant moves slowly (i.e., the minimizers c…
RANDomized-exploration policy Optimization via Multiple Importance Sampling with Truncation (RANDOMIST) for PO with mediator feedback.
problem Policy Optimization in continuous control tasks.
method RANDomized-exploration policy Optimization via Multiple Importance Sampling with Truncation (RANDOMIST) for regret minimization in PO.
result Achieving constant regret under certain circumstances in PO with mediator feedback.
Capacity-Constrained Online Convex Optimization with Delayed Feedback
problem Online learning with delayed feedback under a hard capacity constraint
method Reduction to a delayed and weighted OCO problem using a scheduler
result First regret guarantees for capacity-constrained OCO under convex and strongly convex losses
One-bit feedback suffices for a bandit problem's optimal strategy.
problem Optimal strategy for multi-armed bandit problem with limited feedback.
method Coding and decoding schemes for one-bit feedback to mimic full-reward feedback.
result Regret ratio approaches 1 with one-bit feedback.
New framework for resilient bi-criteria optimization under noisy feedback.
problem Bi-criteria combinatorial optimization with noisy function evaluations.
method Introducing (α,β,δ,extttN)-resilience and developing a black-box framework. result Achieves sublinear regret and constraint violation for bi-criteria bandit problems.
We use online convex optimization (OCO) for setpoint tracking with uncertain, flexible loads. We consider full feedback from the loads, bandit feedback, and two intermediate types of feedback: partial bandit where a subset of the loads are individually observed and the rest are observed in aggregate, and Bernoulli feed…
The paper tackles combinatorial pure exploration with various feedback structures and proposes efficient algorithms.
problem Identifying the optimal action in a combinatorial space with limited feedback and nonlinear rewards.
method Designs polynomial-time adaptive algorithms for CPE-BL and CPE-PL, providing sample complexity analyses.
result The proposed algorithms achieve sample complexity close to lower bounds and outperform existing methods.
Neural algorithms optimize arm selection with human preference feedback for complex reward functions.
problem Optimizing arm selection with noisy human preference feedback for complex, non-linear reward functions.
method Neural network to estimate reward function using preference feedback, upper confidence bound and Thompson sampling algorithms.
result Sub-linear regret guarantees for efficient arm selection in contextual dueling bandits.
New algorithm allows IGL to work with action-inclusive feedback.
problem IGL's failure in scenarios with action-inclusive feedback.
method Developed an algorithm and provided theoretical guarantees.
result Demonstrated effectiveness on large-scale experiments.
RLHF uses human feedback to train AI models, posing statistical challenges.
problem Aligning AI models with human preferences using noisy, subjective feedback.
method Supervised fine-tuning, reward modeling, policy optimization, statistical ideas.
result Statistical methods for reward function learning and policy optimization.
Combines BO with context to optimize binary feedback.
problem Optimizing expensive binary functions with context.
method Bayesian active learning and optimization.
result Efficiently chooses best context and parameters.
MAXMINLCB optimizes unknown target functions with preference feedback using a Stackelberg game approach.
problem Optimizing unknown target functions with pairwise comparisons and human feedback.
method MAXMINLCB, a zero-sum Stackelberg game, balances exploration and exploitation.
result MAXMINLCB consistently outperforms existing algorithms with a rate-optimal regret guarantee.
Proposes a model to optimize feedback for content creators on social media.
problem Nurturing content creation on social media platforms.
method Modeling feedback distribution to incentivize creators and optimize newsfeed experience.
result Optimized newsfeed for content creators without compromising user experience.
Dual active learning improves RLHF by selecting optimal conversations and teachers.
problem Efficiently aligning LLMs with human preferences using RLHF from feedback.
method Offline RL for conversation and teacher selection, dual active reward learning, pessimistic RL.
result The proposed algorithm achieves minimal generalized variance and outperforms state-of-the-arts.
New method tackles composite optimization with error feedback.
problem Challenges in distributed machine learning training and message compression.
method Combines Dual Averaging with EControl for composite optimization.
result First strong convergence analysis for composite optimization with error feedback.
New method learns from subgroup feedback in complex systems.
problem Optimizing complex systems with heterogeneous components.
method Decomposed Gaussian Process (GP) regression and optimization algorithm.
result Proved lower variance and improved accuracy in subgroup feedback.
New algorithm for active learning from feedback coding.
problem Efficiently selecting examples for labeling in active machine learning.
method Formalized structural similarities between active learning and feedback channel coding, developed an optimal transport-based algorithm called Approximate Posterior Matching (APM).
result Learning performance comparable to existing methods at reduced computational cost.
We consider wireless transmission of images in the presence of channel output feedback. From a Shannon theoretic perspective feedback does not improve the asymptotic end-to-end performance, and separate source coding followed by capacity-achieving channel coding, which ignores the feedback signal, achieves the optimal …
Continuous-time distributed mirror descent with integral feedback converges to global optimum.
problem Distributed optimization of a global strongly convex function with local convex components.
method Continuous-time distributed mirror descent with integral feedback.
result Asymptotic convergence to global optimum with constant step-size.
New algorithm for recommending best arms with aggregated feedback.
problem Finding the best arm under aggregated feedback when precise rewards are unavailable.
method Gaussian Process Optimistic Optimisation (GPOO) algorithm with adaptive tree construction.
result The proposed algorithm achieves new simple regret bounds with aggregated feedback.
New algorithm controls linear systems with bandit feedback, achieving optimal regret.
problem Controlling linear systems with bandit feedback under adversarial costs.
method Developed a new algorithm for linear control with memory optimization technique.
result Achieved optimal regret growth proportional to square root of time horizon.
Stochastic structured prediction under bandit feedback follows a learning protocol where on each of a sequence of iterations, the learner receives an input, predicts an output structure, and receives partial feedback in form of a task loss evaluation of the predicted structure. We present applications of this learning …
Sayer uses implicit feedback to optimize system policies.
problem Leveraging implicit feedback to improve system policies is difficult due to bias and incompleteness.
method Sayer combines randomized exploration and unbiased counterfactual estimators to evaluate and train new policies using implicit feedback.
result Sayer can accurately evaluate and train new policies that outperform existing ones.
We propose a generalization of the best arm identification problem in stochastic multi-armed bandits (MAB) to the setting where every pull of an arm is associated with delayed feedback. The delay in feedback increases the effective sample complexity of standard algorithms, but can be offset if we have access to partial…
Unified LP framework for offline reward learning from human demonstrations and feedback.
problem Reward learning from human demonstrations and feedback with robustness and sample efficiency.
method A novel linear programming framework for offline reward learning.
result Unified LP framework achieves better performance compared to MLE.