arXiv subjects

Ravi Kumar

Publications and source records attributed to Ravi Kumar.

At least 19 recordsLinked to original sources

arXiv2026

VeriCUA: Verification as Agentic Exploration for Computer-Using Agents

Building and scaling computer-using agents (CUAs) across desktop, mobile, and web environments requires reliable task-completion verification for data curation, evaluation, and reinforcement learning. Typical automated approaches prompt vision-language models (VLMs) to judge trajectories or generate programs to inspect application state. However, both rely on partial views of complex digital environments. VLM judges are restricted to supplied observations, while generated programs depend on accessible state. These views can omit decisive evidence or leave completion judgments unsupported. Tailoring separate verifiers to different platforms and access settings adds further engineering cost. We present VeriCUA, a platform-agnostic framework for verification through agentic exploration. It treats verification as dynamic evidence gathering and assessment within a shared procedure for online and offline settings across platforms and observation modalities. After inspecting the available context to construct a task rubric, the verifier selects GUI or code checks for unresolved requirements and combines their findings under explicit evidence rules. Experiments show improved agreement with reference labels over the compared baselines across these settings, with the largest offline gains on hard cases and stronger rejection of false completion claims online. Further analyses connect these gains to complementary evidence and test-time exploration.

cs.AI↗

arXiv2026

Photon-energy-programmable subnanometric electron birth-site control

Optical control of electron-generation sites has broadly enabled ultrafast nanoscale imaging, spectroscopy, and functional control. Existing approaches achieve nanoscale site selectivity by shaping localised optical fields around nanostructures, thereby limiting independent site selectivity within the same local-field hotspot. Here, using a single-molecule electron emitter, we show that site selectivity can instead be encoded in the electronic excitation pathway, enabling subnanometric control of electron birth sites within a common local-field hotspot. By tuning the photon energy, we selectively access different groups of molecular states whose combined contributions define emission channels with distinct spatial character, thereby reversibly switching the electron birth site within the same emitter. The switching is read out directly in the far-field emission pattern and occurs upon changing the photon energy, but not upon varying the intensity or polarisation. Our results establish optical birth-site selectivity that is not dictated by the local-field distribution, opening a route to electron birth-site control through the electronic excitation pathway.

cond-mat.mes-hall↗

arXiv2026

Bouba Beats Kiki, and Other Regularities in the Subjective Preferences of LLMs

Users engage with large language models on tasks that are purely objective ("When was Netflix founded?"), purely subjective ("What should I name my child?") or anywhere in-between. We have rich tools for evaluating model accuracy with respect to objective criteria. However, when it comes to subjective aspects of a task, we do not have similarly developed methodology at hand. In this paper, we take some early steps towards framing the question of subjective evaluation. Rather than measure how often a model is right or wrong on a question, we instead ask, how do models make choices among multiple viable responses without a clear winner? To study this question, we restrict our focus to short-form content, and consider comparison tasks among three such content types: numbers, baby names, and fictional cities. We identify 29 aspects of these short-form examples covering many areas of syntax and semantics (length, alphabetic order, popularity, vowel mix, spoken sound, etc.). We show first that models tend to have strong and reproducible preferences for or against some of these aspects. Second, we show that variations in how the model is asked to express a preference may cause some of these aspects to activate or deactivate in the model's selection. Third, we show that overall preference distributions among pairs may vary radically based on seemingly-unimportant variations in the nature of the question posed to the model, raising new questions about manipulation. And finally, we show that, if models are used both to create lists of alternatives and to rank them, there are unexpected interactions between the two stages that may also drive the models' stated preferences.

cs.CL↗

arXiv2026

CALLIOPE: Private Insights into AI Use, Without Embeddings

We introduce CALLIOPE, a top-down framework for discovering topic hierarchies in large corpora of unstructured data (e.g., LLM conversations), while supporting differential privacy (DP). Unlike traditional and even recent LLM-augmented approaches that ultimately rely on clustering vector embeddings, our framework leverages LLM capabilities directly for the entire pipeline. Experiments demonstrate that this approach provides better utility over existing (embedding-based) methods in the DP setting, especially in the high privacy regime. We evaluate the quality of CALLIOPE outputs using a variety of traditionally studied clustering metrics, as well as some novel utility metrics that we introduce.

cs.LG↗

arXiv2026

Orbitally Tuned Composite-Fermion Pairing Transitions in Bilayer Graphene

The effective interaction between composite fermions, set entirely by the Coulomb potential and the underlying electronic Landau level orbitals, can stabilize exotic fractional quantum Hall states. In particular, half-filled Landau levels with different orbital character can host either metallic or paired phases of composite fermions. Here, we leverage experimental control over the orbital composition to realize a composite-fermion pairing transition in the first excited Landau level of bilayer graphene. Transport measurements at filling factors ν = 9/2 and 11/2 reveal conductive states giving way to well-developed plateaus with increasing displacement fields. These states are insensitive to an in-plane magnetic field, indicating single-component ground states and thus pointing at non-Abelian orders. Our numerical study, based on displacement-field-dependent Landau-level wave functions, supports the orbital origin of the pairing transition and suggests Moore-Read or anti-Pfaffian ground states.

cond-mat.mes-hall↗

arXiv2026

A General Framework for Matching with Delays

We present a simple deterministic $O(\log n)$-competitive algorithm for the online minimum-cost perfect matching with delays (MPMD) problem on arbitrary, unknown metric spaces, where $n$ is the (unknown) number of requests. This improves upon the recent $O(\log^5 n)$-competitive algorithm of Dufay and Wattenhofer (SODA 2026) and nearly matches the $Ω(\log n / \log \log n)$ randomized lower bound of Ashlagi et al. (APPROX 2017). Our algorithmic framework rests on two ingredients: an online moat-growing process that grows active dual variables at a deliberately slowed constant rate $α< 1$ to control temporal asymmetry, and a delayed eligibility rule that matches requests in the same dual component only after their waiting time is proportional to their time-expanded distance. The framework extends naturally to arbitrary concave delay functions, maintaining an $O(\log n)$ competitive ratio, and to $k$-way matching with delays, yielding an $O(k \log n)$-competitive deterministic algorithm.

cs.DS↗

arXiv2026

Asymptotic Behavior of Transformers

In this paper we study the behavior of transformer networks generating text over long periods of time. We show that even a three-token, single-layer attention-only transformer can fail to converge. We then study a rich class of transformers with arbitrary finite depth, layer normalization, MLPs, RoPE encodings, and an arbitrary mix of local and global layers. Every transformer in this class has a softmax temperature threshold above which its empirical cache distribution converges to a unique limit, independent of initial conditions. We show further that, while each transformer has a temperature threshold, the supremum of this set is infinite, so there is no global minimal temperature that forces all transformers to converge.

cs.LG↗

arXiv2026

Sample-Efficient Optimization over Generative Priors via Coarse Learnability

We study zeroth-order optimization where solutions must minimize a cost $d(s)$ while maintaining high probability under a complex generative prior $L(s)$ (e.g., a parameterized model). This reduces to sampling from a target distribution proportional to $L(s) e^{-T \cdot d(s)}$. Since classical model-based optimization (MBO) lacks finite-sample guarantees for expressive approximate learners, we introduce "coarse learnability", a flexible statistical assumption requiring only that a learned model covers the target's probability mass within a polynomial factor. Leveraging this assumption, we design an iterative MBO algorithm called \alift with a sample correction step that provably approximates the target using only a polynomial number of samples. We apply this framework to globally optimizing non-convex objectives bounded by a quadratic envelope in $R^n$, where we show this assumption is naturally satisfied for a family of "optimistic" posterior distributions. To reach global $\varepsilon$-optimality, this implies a sample complexity of $\widetilde{O}(\log 1/\varepsilon)$, a rate characteristic of optimistic space-partitioning methods. We further justify coarse learnability as an assumption for generative priors theoretically, proving that in simple settings, parametric maximum likelihood estimation and over-smoothed kernel density estimators naturally satisfy it. Finally, one motivation for our framework comes from inference-time alignment. Though our primary contribution pertains to the theoretical foundations of MBO, we provide qualitative evidence that, in simple settings, even primitive LLMs can shift their distributions toward lower-cost regions when fine-tuned with zeroth-order feedback.

cs.LG↗

arXiv2026

Online Algorithms with a Sample: Tight Bounds and Adversarial Robustness

Suppose an online algorithm is given an unbiased $p$-sample of its input as offline advice; can the algorithm exploit the sample to achieve beyond-worst-case performance? We study this online algorithms with a sample (OAS) model. We show a tight $O\left(\log (1/p) \cdot \log m + \log n\right)$-competitive algorithm for set cover, exponentially improving upon the $O\left(1/p \cdot \log (mn)\right)$ guarantee of Gupta et al. (SODA'24) and answering an open question therein. Our techniques extend to covering integer programs and non-metric facility location, also yielding tight bounds for these problems. Further, we give an $O(\log (1/p)/ \log \log (1/p))$-competitive algorithm for metric facility location, answering an open question of Argue et al. (NeurIPS'22). We then introduce and study the robust variant of the OAS model, in which an adversary is allowed to arbitrarily modify $k$ elements of the $p$-sample. For set cover, covering integer programs, and non-metric facility location, we obtain a tight competitive ratio of $O\left(\log (k/p) \cdot \log m + \log n\right)$. For metric facility location and Steiner tree, we obtain tight competitive ratios of $O\left(\log (k/p) / \log \log (k/p) \right)$ and $O\left(\log (k/p)\right)$ respectively. To the best of our knowledge, these are the first results for robust algorithms in the OAS setting.

cs.DS↗

arXiv2026

Differential Privacy Meets Fixed Parameter Tractability: Algorithms and Lower Bounds

We study combinatorial optimization problems under the constraint of $ε$-differential privacy ($ε$-DP). Given the strong lower bounds for explicitly outputting solutions, we work within the implicit representation framework of Gupta et al. (SODA 2010), where a private polynomial-time randomized "encoder" generates a representation of a solution, and a "decoder" uses this representation along with the input to extract a valid final solution. In this work, we generalize this framework by allowing the encoder to run in fixed-parameter tractable time. This circumvents approximation barriers inherent to polynomial-time algorithms and obtains improved guarantees for many fundamental combinatorial optimization problems. Finally, we establish the first representation-independent lower bounds for our framework. Assuming a non-uniform variant of the Gap Exponential Time Hypothesis, for sufficiently small $ε> 0$, we prove that no $ε$-DP encoder-decoder pair can achieve certain approximation guarantees, if the decoder runs in subexponential time. We further provide representation-dependent lower bounds that hold even for larger $ε$.

cs.DS↗

arXiv2026

Overcoming the Randomness-Utility Trade-off in Answering Differentially Private Linear Queries

We study the question of answering linear queries with differential privacy using few (expected) random bits. We provide a randomness-efficient analog of the $\| \cdot \|_K$-norm mechanism of Hardt and Talwar [HT10]. For the $\ell_\infty$-error, our algorithm can answer $d$ linear queries with $O(d / \varepsilon)$ error using $O(\log d)$ random bits, improving upon algorithms of Canonne et al. and Ghentiyala [CSV25, Ghe26]; this is optimal when $\varepsilon \le 1/d$. We also provide a computationally efficient version of our algorithm, albeit with an $O(\log d)$ multiplicative increase in the error.

cs.CR↗

arXiv2026

Online Differentially Private Consistent Clustering

We study differentially private (DP) $k$-means and $k$-median clustering in the online streaming setting. In this model, points arrive sequentially, and at each time step, we need to output a set of $k$ centers that optimizes the clustering objective for all points seen so far. We give a generic reduction that transforms the (sensitive) input stream into a private stream, which is a semi-coreset of the input stream. This implies that any (non-private) online clustering algorithm, run as a post-processing step, can achieve good utility for the original clustering objective. Our algorithm matches or improves upon the approximation ratio, space usage, and running time of existing algorithms [Epasto et al., 2026, Dupré la Tour et al., 2024]. A key aspect of our reduction is that it inherits desirable properties of the underlying non-private clustering algorithm, such as consistency [Lattanzi and Vassilvitskii, 2017]--a property not satisfied by previous DP algorithms.

cs.DS↗

arXiv2026

Multi-Level Aggregation via Dual Fitting: An $O(D)$-Competitive Algorithm

We present a new online algorithm for the well-known Multi-Level Aggregation Problem (MLAP) with arbitrary delay functions, achieving a $2D$-competitive ratio, where $D$ is the depth of the underlying tree. This result improves the current best-known competitive ratio of $O(D^2)$ and asymptotically matches the $D$-competitive bound previously known only for the deadline variant, thereby closing the asymptotic gap between the two settings. Our key technical contribution is a novel dual fitting framework that provides a unified analysis for both settings; in particular, it also establishes a $D$-competitive ratio for MLAP with deadlines. Our analysis is built upon two new ideas: a hindsight dual construction, which resolves the infeasibility issues in traditional online primal-dual methods, and a time-dependent dual packing that maintains feasibility over dynamic request sets.

cs.DS↗

arXiv2026

A Matrix Factorization Approach in Turnstile Streaming

We define the $M$-point query problem in data streams. Given a fixed matrix $M$, the goal is to maintain a vector $x$ under turnstile updates and answer each query $u$ with an estimate $\widehat{y}_u$ satisfying $|y_u-\widehat{y}_u| \leq \varepsilon \|x\|_1$, where $y=Mx$. We show that if $M$ admits a factorization $M=AB$, where $A$ and $B$ have space-efficient representations, then there is a streaming algorithm using $O(\varepsilon^{-1}\|A\|_{2\rightarrow\infty}\|B\|_{1\rightarrow 1}+(\varepsilon^{-1}\|A\|_{\infty\rightarrow\infty}\|B\|_{1\rightarrow 1})^{2/3})$ words of memory. An important special case is the lower-triangular all-ones matrix, which corresponds to the quantiles problem with additive error $\pm \varepsilon n$, where $n$ is the database size. Our framework generalizes the dyadic approach of Cormode and Muthukrishnan (J. Algorithms, 2005) for turnstile quantiles, and simplifies and improves the analysis of the state-of-the-art dyadic CountSketch algorithms of Wang et al. (SIGMOD, 2013) and Luo et al. (VLDB, 2016). Our approach is also related to the matrix mechanism of Li et al. (VLDB J., 2015) in differential privacy: given a database $x\in\mathbb{R}^U$ and a matrix $M$, the mechanism outputs a private approximation to $Mx$, with the privacy-error tradeoff governed by a matrix factorization norm of $M$. We also improve the prior lower bound for quantiles with deletions, showing a memory lower bound of $Ω(\varepsilon^{-1}\log U)$ words. We also show any factorization has $\|A\|_{2\rightarrow\infty}\|B\|_{1\rightarrow 1} = Ω((\log^{1.5} U) / \log\log U)$. This lower bound is new, and shows that for quantiles, the dyadic CountSketch is nearly optimal amongst factorization-based approaches.

cs.DS↗

arXiv2026

Van Hove singularity-driven giant Nernst signal in twisted double bilayer graphene

Twisted graphene layers host van Hove singularities (vHSs), peaks in the electronic density of states, thought to drive exotic phases in moiré materials, but their effect on thermal transport has remained unclear. Here we show that vHSs in twisted double bilayer graphene (tDBLG) generate an unusually large Nernst signal-the transverse voltage produced by a longitudinal temperature gradient in a magnetic field. The pronounced Nernst peaks at the vHSs of the conduction and valence bands of tDBLG are tunable by an electric field with a maximum value of $\sim 40$ $μV K^{-1} T^{-1}$ at $\sim 1$ $K$, which is comparable to the best-known Nernst materials. Our theoretical calculations show that the large enhancement of the Nernst signal arises from the Lifshitz transitions around the vHSs. These findings establish the Nernst effect as a sensitive probe of Fermi-surface topology in moiré materials, and identify a universal thermoelectric signature of van Hove singularities.

cond-mat.mes-hall↗

arXiv2026

Fixed-Parameter Tractability of Private Synthetic Data Generation

We study the problem of generating synthetic data under differential privacy. We establish fixed-parameter tractability (FPT) for this problem where the parameter is the treewidth of the query family's incidence graph. Our algorithms attain optimal error rates across all regimes and are realized by two different approaches: the first is based on linear programming (LP) and the FPT of the separation problem for the LP dual; the second is based on a subsampled private multiplicative weights method, where we obtain FPT for sampling from Gibbs distributions. Both approaches are unified by a dynamic programming framework over a tree decomposition.

cs.DS↗

arXiv2026

Stochastic Caching via Subset Entropy

A classic approach to beyond worst-case algorithm design is to impose stochastic assumptions on the input. However, a limiting feature of stochastic analyses is that, by the min-max principle, performance on worst-case distributions mirrors that of randomized algorithms on worst-case inputs. In other words, the same shortcoming of worst-case analysis -- its inability to distinguish "easy" and "hard" instances -- reappears as an inability to distinguish "easy" and "hard" distributions. This raises a natural question: Can we characterize "easy" input distributions with useful beyond worst-case bounds? A canonical example is the stochastic caching problem (Aho et al. 1971). When the page requests are drawn i.i.d. from the uniform distribution, the best achievable competitive ratio is $O(\log k)$, matching the performance of the best randomized algorithm on worst-case instances (Fiat et al. 1991). However, when the input distribution has less entropy, intuition suggests that we should be able to do better by exploiting the information provided by the distribution. We formalize this by defining a new information-theoretic parameter called subset entropy which we use to give a fine-grained characterization of the competitive ratio of stochastic caching, including a new analysis for the well-known LRU algorithm on stochastic inputs. While our technical results are for the caching problem, we believe the broader principle -- parameterizing algorithmic performance by an entropy measure of the input -- is of independent interest and might apply to other online/stochastic optimization problems. Indeed, for problems such as (comparison-based) sorting, online matching, load balancing, etc., the hardest stochastic instances involve high-entropy distributions. We hope our work is a step toward a broader theory of fine-grained algorithmic performance for this class of problems.

cs.DS↗

arXiv2026

Adaptive Weighted Averaging

We study the problem of selecting the largest among $n$ unknown values $x_1,\dots,x_n$ given only a single unbiased estimate $y_i$ for each $x_i$. We design strategies that are simultaneously admissible (not uniformly dominated by any other strategy) and also never worse than a given baseline such as uniform random selection. We provide an application to stochastic optimization, where we obtain online-to-batch conversion bounds with a desirable "no-compromise" guarantee: they are never worse than standard random iterate selection, and yet can be significantly better in benign settings.

cs.LG↗