Research
On-device research index

arXiv research

A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.

168,738 papers · 148 categories

Trend · papers per month

4487131174 · Jun 202019922001200920172026
48 results for online alignment

SAIL improves online alignment of large language models with minimal feedback.

problem Offline RLHF methods often lead to sub-optimal performance due to fixed preference datasets.
method SAIL uses bilevel optimization and a single-level first-order method to iteratively refine model alignment.
result SAIL significantly improves alignment performance on open-sourced datasets with minimal computational overhead.

Paper addresses online alignment of large language models under uncertain preference feedback.

problem Online alignment of large language models with misspecified preference feedback.
method Formulates an oracle-robust objective as a worst-case optimization problem for log-linear policies, and develops projected stochastic composite updates.
result Shows that the robust objective admits an exact closed-form decomposition and achieves O~(ε2)\widetilde{O}(\varepsilon^{-2}) oracle complexity.

GALA adapts learning rates online by aligning gradients, improving deep learning model performance.

problem Fine-tuning learning rates for deep learning models requires extensive grid search.
method GALA dynamically adjusts learning rates by tracking gradient alignment and local curvature.
result GALA produces a flexible, adaptive learning rate schedule that increases when gradients align.

Paper shows equivalence between two alignment methods and introduces a new algorithm.

problem Ensuring human alignment of large language models for useful, safe, and pleasant user experience.
method Introduces IPO-MD algorithm, showing equivalence between IPO and Nash-MD methods.
result Equivalence between IPO and Nash-MD methods proven when considering online version of IPO.

One-step Bellman alignment improves online RL by reducing task mismatch.

problem Online RL struggles with task similarity defined by rewards or transitions.
method One-step Bellman alignment and re-weighted targeting (RWT) to correct task mismatch.
result Regret bounds show task shift complexity, not target MDP, affects performance.

A3RL combines online and offline RL with active sampling to improve policy learning.

problem Combining online and offline RL for sample efficiency and robustness.
method A3RL uses a confidence-aware Active Advantage Aligned (A3) sampling strategy to prioritize data from both online and offline sources.
result A3RL outperforms competing online RL techniques that use offline data.

DRO-REBEL improves LLM alignment by robustly updating models online.

problem Overfitting and drifting of LLMs during RLHF.
method DRO-REBEL uses type-pp Wasserstein, KL, and χ2χ^2 ambiguity sets for robust online updates.
result DRO-REBEL achieves faster convergence and better performance than prior methods.

An efficient algorithm for aligning diffusion trees to networks with information asymmetry.

problem Aligning diffusion trees to networks with information asymmetry.
method Tree correlation tests for extracting alignment information.
result Explicit lower bounds on the probability of correct matches for each vertex on the diffusion tree.

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.

Optimal hidden-target learning for online inventory optimization on general convex sets.

problem Online inventory optimization (OIO) on arbitrary bounded convex capacity sets.
method Maintaining a hidden target and projecting it onto the feasible order-up-to set.
result The method improves the best known regret guarantee for OIO on general convex sets from inverse to inverse-square-root dependence on the common-demand probability.

Online BSP-Forest improves space partitioning for large-scale classification and regression.

problem Efficient space partitioning for large-scale classification and regression problems.
method Developed an online BSP-Forest framework that expands space coverage and refines partition structure in real-time.
result Guaranteed universal consistency for both classification and regression problems.

Adversarial online nonparametric regression achieves optimal rates with locally adaptive learning.

problem Adversarial online nonparametric regression with general convex losses.
method Parameter-free learning algorithm leveraging chaining trees to compete against H{ö}lder functions, dynamically tracking and adapting to local smoothness variations.
result First computationally efficient algorithm with locally adaptive optimal rates for online regression in an adversarial setting.

Paper addresses RLHF alignment challenges with novel algorithms.

problem Challenges in RLHF alignment, especially in strategic exploration.
method Develops a reverse-KL regularized contextual bandit formulation and proposes efficient algorithms with theoretical guarantees.
result Proposed methods significantly outperform existing RLHF algorithms in real-world experiments.

Kolmogorov-Arnold Networks enable ultrafast online learning with fixed-point quantization.

problem Efficient online learning for high-frequency systems with strict memory constraints.
method Fixed-point online training on FPGAs exploiting B-spline locality in KANs.
result Kolmogorov-Arnold Networks are more efficient and expressive than MLPs for low-latency tasks.

SARA uses similarity to learn rewards robustly and adaptively.

problem Robustness to labeler errors and adaptability to diverse feedback formats.
method Contrastive framework that learns latent representations and computes rewards as similarities.
result Strong performance on offline RL benchmarks and diverse applications.

Efficiently identifies good policies by choosing contexts for human feedback.

problem Efficiently acquiring human feedback for preference alignment in large language models.
method Formalizes active exploration as a dueling bandit problem and proposes an active exploration algorithm with a polynomial worst-case regret bound.
result Proposed method outperforms baselines with limited human preferences on various language models and datasets.

Paper resolves bias in ALFT training using generalized alignment games.

problem Systematic bias in estimating logarithmic rewards from small batches.
method Generalized Distributional Alignment Games, U-statistics, minimax polynomial estimators, Variance-Optimal Augmented Polynomial Optimization Program (AQP) Estimator.
result Proves optimal bias and accelerated convergence in ALFT training.

SAIL-RevKL improves SAIL's convergence by regularizing the objective function.

problem Convergence of self-improving online LLM alignment algorithms.
method Proposed SAIL-RevKL, a regularized objective function to improve optimization landscape.
result Proved SAIL-RevKL satisfies the Polyak-Lojasiewicz (PL) condition with near-linear sample complexity.

Paper tackles entity matching over multi-source data, optimizing alignment and mitigating negative transfer.

problem Learning effective entity matching models over multi-source large-scale data with relaxed assumptions.
method Proposes a Relaxed Multi-source Large-scale Entity-matching (RMLE) problem and Incentive Compatible Pareto Alignment (ICPA) method.
result Optimized cross-source alignments and mitigated negative transfer, improving entity matching accuracy.

OSAMD adapts online to changing distributions with limited labels.

problem Models struggle with continual distribution shifts and expensive labeling in changing environments.
method Online Active Continual Adaptation with OSAMD, an online teacher-student structure and margin-based criterion.
result OSAMD achieves favorable dynamic regret bounds under changing environments with limited labels.

Paper proposes faster adaptation to distribution shifts in online settings.

problem Violation of exchangeability assumption in evolving data environments.
method Online conformal inference with retrospective adjustment.
result Faster adaptation to distributional shifts demonstrated through numerical studies.

Having a sequence-to-sequence model which can operate in an online fashion is important for streaming applications such as Voice Search. Neural transducer is a streaming sequence-to-sequence model, but has shown a significant degradation in performance compared to non-streaming models such as Listen, Attend and Spell (…

2017-12-05abs ↗pdf ↗

Paper proposes a faster, higher-quality RL method for text-to-image models.

problem Improving image quality and prompt alignment in RL post-training of text-to-image models.
method Online RL variant that reduces variance by sampling paired trajectories and optimizing flow velocity.
result Method converges faster and yields higher output quality and prompt alignment.

This paper proposes an online knowledge distillation method that transfers feature map information in addition to class probabilities.

problem Previous online knowledge distillation methods only utilized class probabilities, missing feature map information.
method Adversarial training framework to transfer feature map information; multiple networks trained simultaneously with discriminators.
result Our method performs better than direct alignment methods and is more suitable for online distillation.

A multi-step framework tackles online unsupervised domain adaptation with novel mean-target subspace computation.

problem Online unsupervised domain adaptation with unlabelled target data arriving sequentially.
method Multi-step framework with a novel mean-target subspace computation and temporal coherency consideration.
result Improved performance over previous approaches on four datasets.

New insights into SGD and SGD-M in high dimensions.

problem Understanding and comparing SGD and SGD-M in high-dimensional settings.
method Developed high-dimensional scaling limits for SGD-M and online SGD, examining their dynamics and performance.
result SGD-M amplifies high-dimensional effects, potentially degrading performance compared to online SGD.

Improves RLHF sample efficiency by scaling reward complexity polynomially.

problem Exponential sample complexity in RLHF algorithms for skewed preferences.
method SE-POPO, an online RLHF algorithm that achieves polynomial sample complexity.
result SE-POPO outperforms existing algorithms in sample efficiency.

Sequence-to-sequence models with soft attention have been successfully applied to a wide variety of problems, but their decoding process incurs a quadratic time and space cost and is inapplicable to real-time sequence transduction. To address these issues, we propose Monotonic Chunkwise Attention (MoChA), which adaptiv…

2017-12-14abs ↗pdf ↗

Proposes a new RL method to fine-tune flow-based models with arbitrary rewards.

problem Challenges in fine-tuning continuous flow-based generative models with arbitrary reward functions.
method Online Reward-Weighted Conditional Flow Matching with Wasserstein-2 Regularization (ORW-CFM-W2)
result Achieves optimal policy convergence with controllable trade-offs between reward maximization and diversity preservation.

Federated online learning for streaming data with privacy and efficiency.

problem Analyzing continuous, heterogeneous data streams in a privacy-preserving manner.
method Personalized models for each data source, subgroup assumption, penalized renewable estimation, proximal gradient descent.
result Effective model for distributed multi-source streaming data analysis with privacy and efficiency.

Recent work in distance metric learning has focused on learning transformations of data that best align with specified pairwise similarity and dissimilarity constraints, often supplied by a human observer. The learned transformations lead to improved retrieval, classification, and clustering algorithms due to the bette…

2017-01-07abs ↗pdf ↗

Online news recommender systems aim to address the information explosion of news and make personalized recommendation for users. In general, news language is highly condensed, full of knowledge entities and common sense. However, existing methods are unaware of such external knowledge and cannot fully discover latent k…

2018-01-25abs ↗pdf ↗

Paper proposes a new RLHF framework for human preference learning.

problem Handling dependent online human preference outcomes with dynamic contexts.
method Two-stage algorithm with εε-greedy followed by exploitation; anti-concentration inequalities and matrix martingale concentration techniques.
result Our method achieves optimal regret bound and asymptotic normality of estimators.

New method for efficient online exploration in RLHF reduces regret.

problem Efficiently collecting new preference data in RLHF to refine reward model and policy.
method Proposes a new exploration scheme that directs preference queries toward reducing uncertainty in reward differences most relevant to policy improvement.
result Establishes regret bounds of order T(β+1)/(β+2)T^{(β+1)/(β+2)} for online RLHF, with polynomial scaling in all model parameters.

New method for RLHF reduces costs by integrating new data in one pass.

problem Continuous integration and re-optimization of models in RLHF leads to high computational and storage costs.
method Proposes a one-pass reward modeling method using online mirror descent with a tailored local norm.
result Achieves constant-time updates per iteration, enhancing both statistical and computational efficiency.

Bayesian rating system for large competitions improves prediction and efficiency.

problem Rating systems for large, competitive events like online programming contests.
method Developed a Bayesian rating system for many participants, proving robustness and runtime.
result The system outperforms existing systems in accuracy and computation speed.

Recent work in distance metric learning has focused on learning transformations of data that best align with provided sets of pairwise similarity and dissimilarity constraints. The learned transformations lead to improved retrieval, classification, and clustering algorithms due to the better adapted distance or similar…

2016-03-11abs ↗pdf ↗

Study improves neural network performance in sequential learning for image classification.

problem Improving neural network performance in sequential learning for image classification.
method Evaluation of approaches for computing prequential description lengths, proposing forward-calibration and replay-streams.
result Improved description lengths for image classification datasets, outperforming previous results.

New approach uses autoregressive models to explore and quantify uncertainty in decision-making.

problem Quantifying and exploring uncertainty in online decision-making.
method Reformulates uncertainty as missing future outcomes, training autoregressive models for next-outcome prediction.
result Establishes a reduction from online learning to offline next-outcome prediction, controlling Bayesian regret by sequence prediction loss.