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.

169,341 papers · 148 categories

Trend · papers per month

106212317423 · Jun 202019922001200920182026
48 results for agnostic bounds

New bounds found for agnostic learning with sample compression schemes.

problem Finding optimal rates of convergence for agnostic learning.
method Established tight characterization of worst-case rates for agnostic learning with sample compression schemes.
result Optimal rates of convergence for size-kk agnostic sample compression schemes are klog(n/k)n\sqrt{\frac{k \log(n/k)}{n}}.

The study optimizes polynomial regression for learning under Gaussian distributions.

problem Agnostic learning of Boolean and real-valued functions under Gaussian distributions.
method LP duality and polynomial degree analysis for L1L^1-regression.
result Optimal SQ lower bounds for various function classes.

New SQ lower bound shows complexity nearly matches known upper bound for smoothed agnostic learning.

problem Smoothed agnostic learning of halfspaces under subgaussian distributions.
method Statistical Query (SQ) lower bound using moment-matching hard distribution and linear programming duality.
result First non-trivial lower bound on complexity nearly matches known upper bound.

New algorithm finds optimal policy with polynomial trajectories in deterministic systems.

problem Finding optimal policy in deterministic systems with function approximation.
method Novel recursion-based algorithm with tight bounds on error and sample complexity.
result Optimal policy found using O(dimE)O(\dim_E) trajectories with $δ= O\left(ρ/\sqrt{\dim_E} ight)$.

New algorithm for reliable learning of Gaussian halfspaces with improved sample and computational complexity.

problem Learning halfspaces under Gaussian marginals with reliable agnostic model.
method Developed a new algorithm for reliable learning of Gaussian halfspaces with specific sample and computational complexity.
result Achieved a new algorithm with improved sample and computational complexity for reliable learning of Gaussian halfspaces.

New algorithm learns disjunctions faster than previous methods.

problem Learning Boolean disjunctions in the agnostic PAC model.
method Developed an agnostic learner with complexity 2ildeO(n1/3)2^{ ilde{O}(n^{1/3})}.
result First separation between SQ and CSQ models in distribution-free agnostic learning.

New method for private density estimation of high-dimensional Gaussian mixtures.

problem Private density estimation for mixtures of unrestricted high-dimensional Gaussians.
method Exploits list global stability to prove upper bound on sample complexity.
result First upper bound on sample complexity for agnostic private density estimation.

Proposes class-agnostic object detection to handle all objects without class labels.

problem Difficulty and cost in creating annotated datasets limit conventional object detection models to specific object types.
method Proposes class-agnostic object detection as a new problem and proposes training and evaluation protocols. Uses adversarial learning to exclude class-specific information.
result Adversarial learning improves class-agnostic detection efficacy.

New algorithm improves active learning in agnostic pool-based classification.

problem Efficient active learning in the agnostic setting with minimized sample complexity.
method Solves an experimental design problem to determine a distribution over examples for label requests.
result Achieves sample complexity bounds never worse than best disagreement coefficient-based bounds, sometimes significantly smaller.

New complexity measure helps in agnostic reinforcement learning with or without access to MDP dynamics.

problem Understanding the number of rounds needed to learn an ε-suboptimal policy in unknown MDPs.
method Introducing spanning capacity as a new complexity measure and developing POPLER algorithm.
result There is a separation between generative and online access models for agnostic learnability.

New bounds show agnostic multiclass learning depends on two dimensions: Natarajan and Daniely-Shalev-Shwartz.

problem Understanding sample complexity in multiclass classification with agnostic learning.
method Developed a novel online procedure based on a self-adaptive multiplicative-weights algorithm.
result Agnostic sample complexity bounds are in the form of DS^(1.5)/ε + Nat/ε^2, nearly tight up to a √DS factor.

New learner achieves optimal agnostic error in small error regime.

problem Optimizing agnostic learning in the small error regime.
method Careful aggregations of ERM classifiers.
result Achieves error $c \cdot τ+ O \left(\sqrt{\frac{τ(d + \log(1 / δ))}{m}} + \frac{d + \log(1 / δ)}{m} ight)$, matching lower bound when τd/mτ\approx d/m.

New research shows logistic regression can achieve optimal error rate for agnostic learning of halfspaces.

problem Agnostic learning of homogeneous halfspaces with logistic loss.
method Constructing a well-behaved distribution and using logistic regression with additional convex optimization steps.
result Logistic regression can achieve Ω(extrmOPT)Ω(\sqrt{ extrm{OPT}}) misclassification risk, matching the upper bound.

New bounds show complex neural networks need many queries to learn.

problem Learning non-polynomial activation functions with Gaussian marginals.
method Gradient boosting procedure to amplify lower bounds on SQ dimension of neural networks.
result Statistical-query lower bounds for ReLU regression with 2ncε2^{n^c} ε queries.

New algorithms save computation in agnostic learning with membership queries.

problem Efficiently learning touchstone classes with membership queries.
method Designing agnostic learning algorithms for circuits with sublinear gates.
result Agnostic learning algorithms for circuits with sublinear gates achieve significant computational savings.

New method learns SIMs with arbitrary monotone activations without strong distributional assumptions.

problem Learning Single-Index Models with arbitrary monotone activations.
method Based on omniprediction with calibrated multiaccuracy and Bregman divergences.
result First agnostic learning result for SIMs with arbitrary monotone activations.

Improved private sample complexity for answering classification queries.

problem Designing an algorithm to accurately answer classification queries while maintaining differential privacy.
method Formally studied in agnostic PAC model, derived new upper bound on private sample complexity.
result Improved private sample complexity bound for answering classification queries.

Query access significantly speeds up learning Multi-Index Models under Gaussian distribution.

problem Agnostically learning Multi-Index Models (MIMs) under Gaussian distribution.
method Query access for MIMs with complexity O(k)poly(1/ε)  poly(d)O(k)^{\mathrm{poly}(1/ε)} \; \mathrm{poly}(d) under standard regularity assumptions.
result Query access gives significant runtime improvements over random examples for agnostically learning MIMs.

Improved private agnostic learning with near-optimal sample complexity.

problem Private agnostic learning with arbitrary privacy parameters.
method Near-optimal sample complexity construction.
result Near-optimal extra sample complexity of \(\widetilde{O}(\mathrm{VC}(\mathcal{C})/α^2)\) for any \(\varepsilon \leq 1\).

Study shows transductive learning is equivalent to PAC learning for most natural loss functions.

problem Understanding the relationship between transductive and PAC learning models.
method Extending existing results and developing new techniques to analyze the equivalence of the two models.
result Transductive learning is essentially equivalent to PAC learning for realizable learning with most natural loss functions.

New algorithms improve label complexity for active multi-distribution learning.

problem Active multi-distribution learning with improved label complexity.
method Developed new algorithms for active multi-distribution learning and established improved label complexity upper and lower bounds.
result Improved label complexity upper and lower bounds for active multi-distribution learning.

Study MAML's generalization in varying tasks, proving bounds on error.

problem Bounding MAML's generalization error across tasks.
method Characterizes MAML's generalization error from two perspectives: recurring and unseen tasks.
result MAML's generalization error depends on the number of tasks and samples per task.

Study robust online learning with adversarial perturbations.

problem Learning robust classifiers in the presence of adversarial perturbations.
method Formulated as an online learning problem, considered both realizable and agnostic learnability, defined new dimension controlling mistake/regret bounds.
result Showed new dimension controls mistake/regret bounds, generalized to multiclass hypothesis classes.

New insights into RL with weak function approximation.

problem Statistical complexity of RL with function approximation in large state spaces.
method Agnostic policy learning framework, exploring environment access, coverage, and representational conditions.
result Characterization of fundamental performance bounds and statistical separations.

Study shows depth improves generalization in deep learning models.

problem Understanding why and when depth improves generalization in deep learning.
method Implementation-agnostic state-transition model to analyze depth and generalization.
result Identifies geometric and semigroup mechanisms that keep entropy contribution saturated or polynomial, clarifying depth's statistical advantage.

Study on learning to predict dynamical systems without assuming their structure.

problem Learning to predict the next state of a dynamical system with unknown evolution function.
method Defined new combinatorial measures to quantify mistake and regret bounds in realizable and agnostic settings.
result In the realizable setting, the number of mistakes can grow arbitrarily with time.

Study agnostic feature-based dynamic pricing models with linear policies and noisy valuations.

problem Tackles dynamic pricing with unknown noise and no assumptions on data.
method Studies two agnostic models: linear policy and linear noisy valuation, presenting algorithms and regret bounds.
result Demonstrates no-regret learning is possible under weak assumptions, but noisy feedback is not significantly more useful than bandit feedback.

New IDS algorithm refines parameter norm bounds for better bandit performance.

problem Frequentist IDS requires tight norm bounds, which are often unavailable in practice.
method Iteratively refines a high-probability upper bound on true parameter norm using data.
result Regret bounds independent of assumed parameter norm, outperforming state-of-the-art algorithms.

New estimator SWITCH improves off-policy evaluation in contextual bandits.

problem Estimating value of a target policy using data from another policy in contextual bandits.
method Proposes SWITCH estimator that uses an existing reward model to outperform IPS and DR.
result Switch estimator achieves better MSE than IPS and DR, often outperforming prior work.

Agnostic federated learning aims to reduce bias in model training across different clients.

problem Federated learning models can be biased towards different clients, leading to unfair outcomes.
method Proposes a new agnostic federated learning framework that optimizes for any target distribution formed by client data.
result The approach naturally leads to fairness in model training across different clients.

New algorithms for linear bandits avoid norm knowledge, reducing regret.

problem Linear bandits require knowledge of norm bound SS on parameter θθ^*, leading to high regret.
method Proposes two novel algorithms for changing and fixed arm sets, analyzing their regret bounds.
result Regret bounds show no significant price for not knowing SS, with no price for fixed arm sets.

Efficiently learns Single-Index Models with constant factor approximation.

problem Learning Single-Index Models under L22L_2^2 loss with unknown link functions.
method An efficient algorithm using alignment sharpness for optimization.
result Achieves constant factor approximation to optimal loss for various distributions and link functions.

The study establishes SQ lower bounds for learning halfspaces and ReLUs under Gaussian marginals.

problem Agnostically learning halfspaces and ReLUs under Gaussian marginals.
method Statistical Query (SQ) lower bounds analysis.
result Proves SQ lower bounds of dpoly(1/ε)d^{\mathrm{poly}(1/ε)} for both problems.

New algorithm for learning functions with bounds on error and sample complexity.

problem Learning [0,1][0,1]-valued functions in a prediction model.
method General-purpose algorithm with upper and lower bounds on expected error and sample complexity.
result Improved bounds on sample complexity and agnostic learning conditions.

NSGD-M optimizes machine learning models without hyperparameter tuning, even under relaxed smoothness.

problem Training machine learning models with optimal complexity under relaxed smoothness assumptions.
method Normalized Stochastic Gradient Descent with Momentum (NSGD-M) without stepsize tuning.
result NSGD-M achieves nearly optimal complexity without prior knowledge of problem parameters.

New algorithm reduces best-in-class regret in contextual bandits.

problem Compete with the best policy in a class without model restrictions.
method Proposes an algorithm that updates policies by minimizing a pessimistic objective, including a clipped inverse-propensity estimate and variance penalty.
result Achieves fast best-in-class regret rates, including polylogarithmic rates in the parametric case.

First proper learning algorithm for Gaussian halfspaces with matching sample and computational complexity.

problem Agnostically learning halfspaces under Gaussian distribution.
method First proper learning algorithm with matching sample and computational complexity.
result First proper learning algorithm for agnostically learning halfspaces under Gaussian distribution with matching sample and computational complexity.