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.

168,742 papers · 148 categories

Trend · papers per month

98196294392 · Jun 202019922001200920172026
48 results for non-realizable case

New DP algorithms achieve near-optimal regret bounds for online learning problems.

problem Online learning problems with zero-loss solutions and differential privacy constraints.
method Developed new Differentially Private algorithms with near-optimal regret bounds.
result Achieved near-optimal regret bounds for various online prediction and convex optimization problems.

P. M. Akhmetiev used a controlled version of the stable Hopf invariant to show that any (continuous) map N -> M between stably parallelizable compact n-manifolds, n\ne 1,2,3,7, is realizable in R^{2n}, i.e. the composition of f with an embedding M\subset R^{2n} is C^0-approximable by embeddings. It has been long believ…

2003-05-12abs ↗pdf ↗

New active learning framework for multiclass classification beyond realizability assumption.

problem Active learning in non-realizable settings with convex model classes.
method Surrogate risk minimization, epoch-based fitting, aggregation of models.
result Achieves label and sample complexity comparable to prior work in non-realizable settings.

We study a recent model of collaborative PAC learning where kk players with kk different tasks collaborate to learn a single classifier that works for all tasks. Previous work showed that when there is a classifier that has very small error on all tasks, there is a collaborative algorithm that finds a single classifi…

2018-05-22abs ↗pdf ↗

New method for distributional off-policy evaluation using Bellman residual minimization.

problem Learning return distribution from offline data generated by a different policy.
method Energy Bellman Residual Minimizer (EBRM) method.
result Established finite-sample error bound for EBRM estimator.

Nielsen realization problem for the mapping class group Mod(Sg)\text{Mod}(S_g) asks whether the natural projection pg:Homeo+(Sg)Mod(Sg)p_g: \text{Homeo}_+(S_g)\to \text{Mod}(S_g) has a section. While all the previous results use torsion elements in an essential way, in this paper, we focus on the much more difficult problem of realization of…

2019-04-20abs ↗pdf ↗

In this paper, we will show that the projection Homeo+(Dn2)Bn\text{Homeo}^+(D^2_n)\to B_n does not have a section; i.e. the braid group BnB_n cannot be geometrically realized as a group of homeomorphisms of a disk fixing the boundary point-wise and nn marked points in the interior as a set. We also give a new proof of a result o…

2018-08-24abs ↗pdf ↗

Study shows certain mapping class groups cannot be realized as subgroup of homeomorphisms.

problem Proving non-realizability of specific mapping class groups.
method Analyzing compactly supported and full mapping class groups of surfaces with genus 3 or order 6 symmetries.
result Proven non-realizability of mapping class groups for surfaces with genus 3 or order 6 symmetries.

Paper tackles MLR prediction error without assuming realizable models.

problem Prediction error in mixture of linear regressions without realizable assumptions.
method Developed algorithms for list-decoding MLR predictions and minimized empirical risk.
result Alternating minimization algorithm finds best fit lines in non-realizable settings.

For every compact surface SS of finite type (possibly with boundary components but without punctures), we show that when nn is sufficiently large there is no lift σσ of the surface braid group Bn(S)B_n(S) to Diff(S,n)\operatorname{Diff}(S,n), the group of C1C^1 diffeomorphisms preserving nn marked points and restricting to t…

2015-06-02abs ↗pdf ↗

This study tightens bounds on how GD and SGD generalize in smooth convex optimization problems.

problem Understanding how GD and SGD generalize in smooth stochastic convex optimization problems.
method Provided tight excess risk lower bounds for GD and SGD under different conditions.
result Lower bounds suggest overfitting occurs and gaps remain in some cases.

We investigate active learning by pairwise similarity over the leaves of trees originating from hierarchical clustering procedures. In the realizable setting, we provide a full characterization of the number of queries needed to achieve perfect reconstruction of the tree cut. In the non-realizable setting, we rely on k…

2019-06-22abs ↗pdf ↗

This paper studies universal rates of ERM for binary classification under agnostic learning.

problem The challenge of achieving universal rates of ERM for binary classification under agnostic learning.
method The paper explores the agnostic universal rates of ERM for binary classification, revealing three possible rates: ene^{-n}, o(n1/2)o(n^{-1/2}), or arbitrarily slow.
result The paper provides a complete characterization of which concept classes fall into each of the three categories of agnostic universal rates.

The study provides error bounds for the generalized Lasso with sub-exponential data.

problem Analyzing the generalized Lasso under sub-exponential data distributions.
method Non-asymptotic analysis using generic chaining-based proof strategy.
result Error bounds for the generalized Lasso can be controlled by two complexity parameters.

Generalizes cohomology ring result for combinatorial line arrangements.

problem Cohomology ring of boundary manifold for combinatorial line arrangements.
method Introduced boundary manifold, constructed homology cycles, computed cohomology ring.
result Cohomology ring of boundary manifold is isomorphic to double of Orlik-Solomon algebra.

Solves open problem on universally consistent online learning with unbounded losses.

problem Open problem on universally consistent online learning with unbounded losses.
method Constructs random measurable partitions of the instance space.
result Simple memorization rule is optimistically universal for any unbounded loss.

To a branched cover between closed, connected and orientable surfaces one associates a "branch datum", which consists of the two surfaces, the total degree d, and the partitions of d given by the collections of local degrees over the branching points. This datum must satisfy the Riemann-Hurwitz formula. A "candidate su…

2010-10-14abs ↗pdf ↗

New bandit algorithm works without realizability assumption.

problem Contextual bandit problems without realizability assumption.
method Computes a constrained regression problem in every epoch, ensuring similar regret guarantees as realizability-based algorithms.
result Ensures similar regret guarantees as realizability-based algorithms, up to a misspecification term.

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.

Algorithm learns arbitrary ReLU neurons under Gaussian inputs.

problem Learn an arbitrary ReLU activation over Gaussian marginals.
method Statistical Query (SQ) algorithm that outputs a ReLU activation achieving O(OPT)+εO(\mathrm{OPT}) + \varepsilon loss.
result First constant factor approximation for arbitrary bias in polynomial time.

New algorithm optimizes beam and rate allocation in mmWave systems for multiple users.

problem Optimizing beam and rate allocation in mmWave systems for multiple users with limited feedback.
method Introducing SAT-CTS, a combinatorial semi-bandit policy with satisficing objective.
result SAT-CTS achieves finite-time regret bounds and reduces satisficing regret in mmWave systems.

This article examines five common misunderstandings about case-study research: (1) Theoretical knowledge is more valuable than practical knowledge; (2) One cannot generalize from a single case, therefore the single case study cannot contribute to scientific development; (3) The case study is most useful for generating …

2013-04-02abs ↗pdf ↗

Worst-Case Sensitivity measures model sensitivity to uncertainty set size.

problem Model sensitivity to uncertainty set size in Distributionally Robust Optimization.
method Introducing Worst-Case Sensitivity as a measure of model sensitivity, and deriving closed-form expressions for various uncertainty sets.
result DRO solutions can be sensitive to the family and size of the uncertainty set, and worst-case sensitivity reflects these properties.

Proposes a new framework for balancing average- and worst-case performance in machine learning.

problem Robustness issues in machine learning, especially in safety-critical domains.
method Probabilistic robustness framework that balances average- and worst-case performance.
result Effective algorithm balances average- and worst-case performance with lower computational cost.

Solves equality case in isoperimetric inequality for non-convex domains.

problem Equality case in relative isoperimetric inequality outside convex sets.
method Analyzes non-convex domains to settle the equality case.
result Solves the equality case for relative isoperimetric inequality outside arbitrary convex sets.

Paper improves worst-case regret bounds for RLSVI in reinforcement learning.

problem Minimizing regret in reinforcement learning with randomized value functions.
method Introduces a clipping variant of Thompson Sampling for RLSVI.
result Achieves a ildeO(H2SAT) ilde{\mathrm{O}}(H^2S\sqrt{AT}) worst-case regret bound.

Options are generally learned by using an inaccurate environment model (or simulator), which contains uncertain model parameters. While there are several methods to learn options that are robust against the uncertainty of model parameters, these methods only consider either the worst case or the average (ordinary) case…

2019-05-22abs ↗pdf ↗

GenAI improves actuarial practices through case studies.

problem Improving actuarial practices using AI.
method Four case studies using LLMs, Retrieval-Augmented Generation, and vision-enabled LLMs.
result GenAI enhances claim cost prediction, market comparisons, and car damage classification.

New algorithms reduce complexity for solving nonconvex optimization problems with stochastic objectives and constraints.

problem Solving nonconvex optimization problems with stochastic objectives and constraints.
method Single-loop quadratic penalty and augmented Lagrangian algorithms with variance reduction techniques.
result Achieved best-known complexity guarantees for solving nonconvex optimization problems with stochastic objectives and constraints.

A method for logistic regression inference using both internal and external data.

problem Inability to estimate intercept and marginal case proportion in case-control logistic regression.
method Empirical likelihood approach integrating internal and external data.
result Intercept parameter becomes identifiable with external information, and all parameters are estimable consistently.

Optimizes bond portfolios to avoid worst-case losses.

problem Finding the worst-case value of a bond portfolio over a range of yield curves and spreads.
method Solves a convex-concave saddle point optimization problem to find the worst-case value and construct a robust portfolio.
result Constructs a bond portfolio that includes the worst-case value, ensuring robustness against market uncertainties.

Study extreme-case Value-at-Risk under IFR distributions, providing guidance for risk management.

problem Understanding extreme-case risk measures under distributional ambiguity and increasing failure rate.
method Characterized extreme-case range Value-at-Risk under mean and variance constraints with increasing failure rate.
result Characterized specific characteristics of extreme-case distributions under IFR constraints.

Qualitative behavior of Bach flow is established on compact four-dimensional locally homogeneous product manifolds. This is achieved by lifting to the homogeneous universal cover and, in most cases, capitalizing on the resultant group structure. The resulting system of ordinary differential equations is carefully analy…

2018-03-21abs ↗pdf ↗

Study evaluates and compares numerical differentiation methods on three case studies.

problem Evaluating and comparing numerical differentiation methods for efficiency.
method Forward, Backward, and Centered Finite-Difference methods applied at two levels of precision.
result Different methods perform differently across case studies, with varying levels of computational cost and accuracy.