arXiv subjects

Sebastian Pokutta

Publications and source records attributed to Sebastian Pokutta.

At least 37 records · Page 2Linked to original sources

arXiv2026

Bounded-Support Additive Latin Transversals

We consider the following additive Latin transversal problem. Given a multiset $A=(a_1,\dots,a_k)$ of elements of $\mathbb Z_m$ and a set $B\subseteq\mathbb Z_m$ of cardinality $k$, the task is to order $B$ as $b_1,\dots,b_k$ so that the sums $a_i+b_i$ are pairwise distinct. When $k=m$, Hall proved that a solution exists if and only if $\sum_{i=1}^m a_i\equiv 0 \pmod m$; moreover, his theorem yields a polynomial-time construction. Alon proved that a solution always exists when $m$ is prime and $k<m$, but no polynomial-time construction is known in general. Our main algorithmic contribution is a direct randomized algorithm for Color-Counted Matching: given an edge-colored graph and prescribed target counts for the colors, find a matching using exactly the prescribed number of edges of each color. If $q$ is the sum of the target counts and $h$ is the number of colors, our base-$(q+1)$ reduction to Exact Red Matching, combined with the algorithm of Mulmuley-Vazirani-Vazirani, gives a randomized algorithm with running time $\left(|V|^2+|E|(q+1)^{h-1}\right)^{O(1)} $ for an input graph $(V,E)$. Thus the dependence on the target matching size is $q^{O(h)}$, up to polynomial factors in the graph size. In contrast, applying the general matching-ILP theorem of Lassota and Ligthart as a black box yields a $q^{O(h^2)}$ dependence for the corresponding fixed-size color-counted instances. Applying this primitive to additive Latin transversals with $s=|\operatorname{supp}(A)|$, we obtain an algorithm in randomized time $(k+\log m)^{O(s)}$. In particular, additive Latin transversals are randomized polynomial-time constructible for every fixed support size.

cs.DS↗

arXiv2026

Joint-Range Inequalities for Nonconvex QCQPs

We study cutting planes for nonconvex quadratically constrained quadratic programs (QCQPs) through a project-then-lift approach inspired by mixed-integer rounding (MIR) inequalities. Given two base valid inequalities for the extended QCQP formulation, we project the associated two-row relaxation into a two-dimensional set and analyze the joint range of quadratic functions in two base inequalities. For the nonconvex joint range, we give a closed-form convex hull description of the projected set; for the convex joint range, we give its semidefinite representation. This yields a new family of joint-range inequalities, which can be lifted back to the extended QCQP formulation. MIR inequalities can handle ``mixed'' terms: continuous variables or fractional linear combinations of integer variables. Similarly, we propose more flexible secant mixed-joint-range inequalities, which better expose and exploit the nonconvex joint range. The proposed approach preserves sparsity, since the support of each lifted inequality is controlled by that of the base inequalities. In preliminary geometric experiments, the joint-range inequalities yield substantial area reduction of the projected relaxation constructed via reformulation-linearization-technique.

math.OC↗

arXiv2026

When Does LLM Orchestration Pay Off? A Controlled Evaluation of Accuracy, Cost, and Task Difficulty

LLM orchestration is often assumed to improve reasoning by allocating additional inference-time computation, yet its gains may not justify its cost. Existing comparisons also frequently overlook differences in optimization effort, making it difficult to isolate the value of orchestration itself. We conduct a controlled evaluation of Self-Refine, Best-of-$N$, and Debate against task-only and chain-of-thought (CoT) single-call baselines across five LLM backbones and three domains: competitive programming, chess puzzles, and mathematics. For comparability, we optimize each method with GEPA under the same optimization budget and evaluate all methods on the same difficulty-stratified benchmark items. Orchestration yields moderate but benchmark-dependent gains: averaged across backbones within each benchmark, the largest improvement is 4.6 percentage points over optimized CoT inference and 4.5 points over task-only inference, while requiring approximately 2 to 4 times the mean total tokens of task-only inference. Human-derived difficulty is associated with lower absolute accuracy in all three benchmarks, but within-benchmark analyses do not indicate that orchestration effects increase with task difficulty. By contrast, exploratory mixed-effects analyses reveal strong interactions between orchestration method and backbone model across all three benchmarks, showing that orchestration effectiveness depends substantially on the underlying model. Our results suggest that orchestration decisions should be model-specific and account for whether moderate accuracy gains justify the additional inference cost. More broadly, evaluations of LLM orchestrations should control optimization effort and report model-specific accuracy--cost trade-offs rather than treating additional inference-time structure as uniformly beneficial.

cs.AI↗

arXiv2026

A Counterexample to Ziegler's Cross-Polytope Conjecture for Simplicial 0/1-Polytopes

Ziegler proved that every simplicial $d$-dimensional $0/1$-polytope has at most $2d$ vertices, and asked whether equality forces the polytope to be centrally symmetric and hence, equivalently, a $0/1$-realization of the $d$-dimensional cross polytope. In this note, we give a negative answer, exhibiting an explicit set of $14$ vertices in $\{0,1\}^7$ whose convex hull is a simplicial $7$-polytope and is not centrally symmetric. Moreover, via exhaustive enumeration we show that up to the symmetries of the cube, there are precisely five such polytopes in dimension $7$ (of two combinatorial types) that are not centrally symmetric.

math.CO↗

arXiv2026

The Weight Distribution of the Third-Order Reed-Muller Code of Length 2048

We compute the weight distribution of the third-order Reed--Muller code RM(3,11) of length 2048. The weight enumerator is assembled from the coset weight enumerators of f+RM(2,10), evaluated for representatives of all 3691560 nonzero GL(10,2)-orbits of Boolean cubic forms in ten variables. The computation rests on a structural theorem: a nondegenerate Boolean cubic form admits a nondegenerate hyperplane restriction, except for a single orbit in each odd dimension. The same pass determines the second-order nonlinearity of every cubic form: the relative covering radius of RM(2,10) in RM(3,10) is 408, attained on 179 orbits. This raises the best known lower bound on the covering radius of RM(2,10) from 400 to 408. A complementary heuristic search shows that the relative covering radius of RM(6,10) in RM(7,10) is at most 32, improving the previous bound of 50.

cs.IT↗

arXiv2026

When Does Sparsity Mitigate the Curse of Depth in LLMs

Recent work has demonstrated the curse of depth in large language models (LLMs), where later layers contribute less to learning and representation than earlier layers. Such under-utilization is linked to the accumulated growth of variance in Pre-Layer Normalization, which can push deep blocks toward near-identity behavior. In this paper, we provide evidence that sparsity-like mechanisms can dampen variance propagation and are associated with improved depth utilization Our investigation covers two sources of sparsity: (i) implicit sparsity, which emerges from training and data conditions, including weight sparsity induced by weight decay and attention sparsity induced by long-context inputs; and (ii) explicit sparsity, which is enforced by architectural design, including key/value-sharing in Grouped-Query Attention and expert-activation sparsity in Mixtureof-Experts. Our claim is thoroughly supported by controlled depth-scaling experiments and targeted layer effectiveness interventions. Across settings, we observe a consistent relationship: mechanisms with reduced effective interaction density tend to exhibit lower output variance and better layer differentiation. We eventually distill our findings into a practical rule-of-thumb recipe for training depth-effective LLMs, yielding a notable 4.6 accuracy improvement on downstream tasks. Our results suggest that sparsity-like design choices are an important and previously underemphasized factor in effective depth scaling for LLMs. Code is available at https://github. com/pUmpKin-Co/SparsityAndCoD.

cs.CL↗

arXiv2026

Classification of Boolean Cubic Forms in Ten Variables

We classify Boolean cubic forms in ten variables up to GL(10,2)-equivalence. The catalog contains all 3691560 nonzero orbits. For every orbit we provide a representative with small monomial count, the stabilizer order, and the alternating rank together with an explicit decomposition. The classification is obtained by rank-stratified enumeration. We verify completeness by the Burnside orbit count and independently by the orbit--stabilizer identity. We also provide a fast, complete GL(10,2)-invariant. By polarization, this gives the first complete classification of alternating trilinear forms in dimension 10 over GF(2).

math.NT↗

arXiv2026

Lower Bounds for Frank-Wolfe on Strongly Convex Sets

We present a constructive lower bound of $Ω(1/\sqrt{\varepsilon})$ for Frank-Wolfe (FW) when both the objective and the constraint set are smooth and strongly convex, showing that the known uniform $\mathcal{O}(1/\sqrt{\varepsilon})$ guarantees in this regime are tight. It is known that under additional assumptions on the position of the optimizer, FW can converge linearly. However, it remained unclear whether strong convexity of the set can yield rates uniformly faster than $\mathcal{O}(1/\sqrt{\varepsilon})$, i.e., irrespective of the position of the optimizer. To investigate this question, we focus on a simple yet representative problem class: minimizing a strongly convex quadratic over the Euclidean unit ball, with the optimizer on the boundary. We analyze the dynamics of FW for this problem in detail and develop a novel computational approach to construct worst-case FW trajectories, which is of independent interest. Guided by these constructions, we develop an analytical proof establishing the lower bound.

math.OC↗

arXiv2026

Neural Concept Verifier: Scaling Prover-Verifier Games via Concept Encodings

While Prover-Verifier Games (PVGs) offer a promising path toward verifiability in nonlinear classification models, they have not yet been applied to complex inputs such as high-dimensional images. Conversely, expressive concept encodings effectively allow to translate such data into interpretable concepts but are often utilised in the context of low-capacity linear predictors. In this work, we push towards real-world verifiability by combining the strengths of both approaches. We introduce Neural Concept Verifier (NCV), a unified framework combining PVGs for formal verifiability with concept encodings to handle complex, high-dimensional inputs in an interpretable way. NCV achieves this by utilizing recent minimally supervised concept discovery models to extract structured concept encodings from raw inputs. A prover then selects a subset of these encodings, which a verifier, implemented as a nonlinear predictor, uses exclusively for decision-making. Our evaluations show that NCV outperforms classic concept-based models and pixel-based PVG classifier baselines on high-dimensional, logically complex datasets and helps mitigate shortcut behavior. Overall, we demonstrate NCV as a promising step toward concept-level, verifiable AI.

cs.LG↗

arXiv2026

Symmetric Extension Complexity of the Spanning Tree Polytope

In this note, we prove a tight lower bound on symmetric extended formulations for the spanning tree polytope of the complete graph. More precisely, let $P_{ST}(K_n)$ be the spanning tree polytope of $K_n$. We show that, for all $n\ge13$, every symmetric extended formulation for $P_{ST}(K_n)$ has at least $\binom n3$ inequalities. Since the classical Martin formulation has a symmetric formulation of size $O(n^3)$, this gives \[ \operatorname{xcs}(P_{ST}(K_n))=Θ(n^3). \]

math.CO↗

arXiv2026

Measurement incompatibility and quantum steering via linear programming

The problem of deciding whether a set of quantum measurements is jointly measurable is known to be equivalent to determining whether a quantum assemblage is unsteerable. This problem can be formulated as a semidefinite program (SDP). However, the number of variables and constraints in such a formulation grows exponentially with the number of measurements, rendering it intractable for large measurement sets. In this work, we circumvent this problem by transforming the SDP into a hierarchy of linear programs that compute upper and lower bounds on the incompatibility robustness with a complexity that grows polynomially in the number of measurements. The hierarchy is guaranteed to converge and it can be applied to arbitrary measurements -- including non-projective POVMs (Positive Operator-Valued Measures) -- in arbitrary dimensions. While convergence becomes impractical in high dimensions, in the case of qubits our method reliably provides accurate upper and lower bounds for the incompatibility robustness of sets with several hundred measurements in a short time using a standard laptop. We also apply our methods to qutrits, obtaining non-trivial upper and lower bounds in scenarios that are otherwise intractable using the standard SDP approach, although such bounds are significantly looser than the ones obtained in the qubit case. Finally, we show how our methods can be used to construct local hidden state models for states (i.e., to prove that a state cannot lead to steering under any possible local measurements), or conversely, to certify that a given state exhibits steering; for two-qubit quantum states, our approach is comparable to, and in some cases outperforms, the current best methods.

quant-ph↗

arXiv2026

Random-Subspace Frank--Wolfe over Strongly Convex Sets

Frank--Wolfe methods avoid projections, but over curved feasible regions the full-space linear minimization oracle (LMO) can itself become the computational bottleneck. We introduce random-subspace Frank--Wolfe (RSFW), the first Frank--Wolfe framework, to our knowledge, that replaces the ambient LMO by exact LMOs over random low-dimensional affine sections of a general feasible set, while preserving feasibility in the original space. For smooth convex objectives over compact strongly convex feasible sets, we prove a dimension-explicit approximate-oracle inequality and derive the standard $O(1/k)$ open-loop rate, with high-probability and almost-sure counterparts. Under short steps and a gradient lower bound, the same geometric control yields linear convergence, and we extend the sublinear theory to finite-sum stochastic gradients. We also show that random sections can improve the local curvature model controlling short steps: for smooth objectives, the quadratic model along a sampled section is governed by the compressed Hessian, yielding computable $d\times d$ curvature constants for quadratic objectives over balls and ellipsoids. These results provide a geometric theory of oracle-side randomization in projection-free optimization.

math.OC↗

arXiv2026

FACET: Multi-Agent AI Supporting Teachers in Scaling Differentiated Learning for Diverse Students

Classrooms are becoming increasingly heterogeneous, comprising learners with diverse performance and motivation levels, language proficiencies, and learning differences such as dyslexia and ADHD. While teachers recognize the need for differentiated instruction, growing workloads create substantial barriers, making differentiated instruction an ideal that is often unrealized in practice. Current AI educational tools, which promise differentiated materials, are predominantly student-facing and performance-centric, ignoring other aspects that shape learning outcomes. We introduce FACET, a teacher-facing multi-agent framework designed to address these gaps by supporting differentiation that accounts for motivation, performance, and learning differences. Developed with educational stakeholders from the outset, the framework coordinates four specialized agents, including learner simulation, diagnostic assessment, material generation, and evaluation within a teacher-in-the-loop design. School principals (N = 30) shaped system requirements through participatory workshops, while in-service K-12 teachers (N = 70) evaluated material quality. Mixed-methods evaluation demonstrates strong perceived value for inclusive differentiation. Practitioners emphasized both the urgent need arising from classroom heterogeneity and the importance of maintaining pedagogical autonomy as a prerequisite for adoption. We discuss implications for future school deployment and outline partnerships for longitudinal classroom implementation.

cs.HC↗

arXiv2026

Efficient Quadratic Corrections for Frank-Wolfe Algorithms

We develop a Frank-Wolfe algorithm with corrective steps, generalizing previous algorithms including blended conditional gradients, blended pairwise conditional gradients, and fully-corrective Frank-Wolfe. For this, we prove tight convergence guarantees together with an optimal face identification property. Furthermore, we propose two highly efficient corrective steps for convex quadratic objectives based on linear optimization or linear system solving, akin to Wolfe's minimum-norm point, and show that they converge in finite time under suitable conditions. Beyond optimization problems that are directly quadratic, we revisit two algorithms - split conditional gradient and second-order conditional gradient sliding - which can leverage quadratic corrections to accelerate their quadratic subproblems. We demonstrate improved convergence rates for the first and broader applicability for the second, which may be of independent interest. Finally, we show substantial computational speedups for Frank-Wolfe-based algorithms with quadratic corrections across the considered problem classes.

math.OC↗

arXiv2026

A Free Lunch in LLM Compression: Revisiting Retraining after Pruning

Post-training pruning can substantially reduce LLM inference costs, but it often degrades quality unless the remaining weights are adapted. Since global retraining is expensive at LLM scale, recent work has largely focused on increasingly sophisticated pruning criteria that aim to select better sparsity patterns without adaptation. We revisit this trade-off through local reconstruction: after pruning, we adapt one subset of the model parameters at a time on a calibration set, training it to match the corresponding intermediate activations of the dense model. We evaluate local reconstruction across model families and scales, up to 72B parameters, and establish three main findings. First, local reconstruction is an effective adaptation mechanism for LLMs: it matches post-pruning retraining while using over an order of magnitude less data and compute, even when using PEFT techniques. Second, reconstruction exhibits a broad "free-lunch" regime in granularity, i.e., the reconstruction parameter window: as long as the reconstructed region contains at least a nonlinear submodule, final quality is largely insensitive to the window size, allowing granularity to be chosen primarily based on memory constraints. In contrast, reconstructing individual matrices, despite being the natural approach often proposed in the literature, consistently underperforms, as small matrix-level errors accumulate into larger activation drift. Lastly, reconstruction reduces the relative importance of the pruning criterion: performance gaps between sophisticated criteria and simple baselines shrink with model scale, making simple methods competitive again. Overall, our results challenge the prevailing view that post-pruning adaptation is impractical for LLMs.

cs.LG↗

arXiv2026

Graph Isomorphism: Mixed-Integer Convex Optimization from First-Order Methods

The graph isomorphism (GI) problem, which asks whether two graphs are structurally identical, occupies a unique position in computational complexity -- it is neither known to be solvable in polynomial time, nor proven to be NP-complete. We propose a convex mixed-integer formulation of the problem and leverage first-order convex optimization to tackle it, following a stream of recent work on optimization-driven graph isomorphism detection. We strengthen our formulation with variable fixing techniques that prove highly effective while preserving the polyhedral structure. We perform extensive computations evaluating the performance of different families of methods including a mixed-integer convex formulation, mixed-integer linear optimization, local search and spectral heuristics over a collection of challenging GI instances. We find that a high level of symmetry is beneficial for optimization-based methods. On the other hand, presolving techniques that detect local substructures to fix variables are crucial for asymmetric instances. The proposed method outperforms the second best approach, the integer feasibility approach, on 6 of the 12 graphs families and is on par with it on symmetric families.

math.OC↗

arXiv2026

Stable Personas: Dual-Assessment of Temporal Stability in LLM-Based Human Simulation

Large Language Models (LLMs) acting as artificial agents offer the potential for scalable behavioral research, yet their validity depends on whether LLMs can maintain stable personas across extended conversations. We address this point using a dual-assessment framework measuring both self-reported characteristics and observer-rated persona expression. Across two experiments testing four persona conditions (default, high, moderate, and low ADHD presentations), seven LLMs, and three semantically equivalent persona prompts, we examine between-conversation stability (3,473 conversations) and within-conversation stability (1,370 conversations and 18 turns). Self-reports remain highly stable both between and within conversations. However, observer ratings reveal a tendency for persona expressions to decline during extended conversations. These findings suggest that persona-instructed LLMs produce stable, persona-aligned self-reports, an important prerequisite for behavioral research, while identifying this regression tendency as a boundary condition for multi-agent social simulation.

cs.HC↗

arXiv2026

FACET: Teacher-Centred LLM-Based Multi-Agent Systems-Towards Personalized Educational Worksheets

The increasing heterogeneity of student populations poses significant challenges for teachers, particularly in mathematics education, where cognitive, motivational, and emotional differences strongly influence learning outcomes. While AI-driven personalization tools have emerged, most remain performance-focused, offering limited support for teachers and neglecting broader pedagogical needs. This paper presents the FACET framework, a teacher-facing, large language model (LLM)-based multi-agent system designed to generate individualized classroom materials that integrate both cognitive and motivational dimensions of learner profiles. The framework comprises three specialized agents: (1) learner agents that simulate diverse profiles incorporating topic proficiency and intrinsic motivation, (2) a teacher agent that adapts instructional content according to didactical principles, and (3) an evaluator agent that provides automated quality assurance. We tested the system using authentic grade 8 mathematics curriculum content and evaluated its feasibility through a) automated agent-based assessment of output quality and b) exploratory feedback from K-12 in-service teachers. Results from ten internal evaluations highlighted high stability and alignment between generated materials and learner profiles, and teacher feedback particularly highlighted structure and suitability of tasks. The findings demonstrate the potential of multi-agent LLM architectures to provide scalable, context-aware personalization in heterogeneous classroom settings, and outline directions for extending the framework to richer learner profiles and real-world classroom trials.

cs.HC↗