Combinatorics
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Monday 14 September
Machine Learning
SWB-DM: A Calibrated Sliced-Wasserstein-Barycenter Aggregator with Delayed-Momentum Caching for Byzantine-Robust Federated Learning under Partial Participation
Robust aggregation methods for federated learning quietly rest on a fragile assumption: that whoever shows up in a given round is a fair sample of the full population. In practice, they rarely are. When only a handful of clients participate per round, even a modest fraction of adversaries can dominate that sample and silently invalidate the finite-sample guarantees that coordinate-wise median, Krum, Bulyan, and trimmed mean all depend on. We introduce SWB-DM to address this directly. SWB treats each slice of a client update as a one-dimensional distribution, computes a trimmed Wasserstein barycenter across clients, and recovers coordinate identity via a medoid-based gauge-fixing step – a heuristic we developed and do not claim it belongs to standard optimal-transport theory. DeMoA-style delayed momentum then caches updates across the full client population each round, decoupling robustness from whoever happened to be sampled. Trim ratio calibration is not cosmetic: under-trimming causes collapse at corruption levels a properly calibrated model survives. Across 448 CIFAR-10 configurations, plus CIFAR-100, FEMNIST, and a 500-client scalability run, we find several mechanistically distinct failure modes. Even-sample coordinate-wise median degrades to a deterministic wrong answer. Krum silently violates its own n greater than 2f+2 precondition and diverges without warning. Bulyan's n greater than or equal to 4f+3 threshold produces a sharp pass/fail boundary. On attacks, IPM defeats order-statistic defenses – including SWB – more reliably than ALIE, confirmed through delta-space measurements against a convergence bound. SWB-DM's cache carries a real warm-up cost, but extending all baselines to the same round budget shows its CIFAR-10 gains are disproportionately large. On CIFAR-100, FLTrust benefits more – for reasons entirely unrelated to caching.
Read the full abstract
Cryptography and Security
Universal Defenses for Tool-Integrated LLM Agents Against Adversarial Attacks
Large Language Model (LLM) agents have demonstrated impressive capabilities across a variety of domains, particularly when integrated with external tools for multi-step task completion. However, they are increasingly vulnerable to adversarial attacks, including direct prompt injection, indirect prompt injection, memory poisoning, and backdoor attacks, which exploit the model's openness to prompt injection and tool manipulation. In this work, we explore practical and generalizable defense strategies within a unified framework across these four attack types. We introduce two universal tool-based defenses: Attacker Tool Filtering, which uses anomaly detection (e.g., Isolation Forest) to identify and remove suspicious tools, and Normal Tool Recalling, a white-box method that restores the agent's original toolset prior to planning. Additionally, we incorporate prompt-based defenses: Chain-of-Thought prompting and self-reflection techniques to enhance reasoning and task paraphrasing to mitigate attacks. Experimental results across both four open-source LLMs (Gemma2-9B, Qwen2-7B, LLaMA3-8B, and LLaMA3.1-8B) and three proprietary LLMs (GPT-3.5, GPT-4, and GPT-5) show that our methods significantly reduce the Attack Success Rates (ASR), achieving 0% ASR in many settings, while preserving or even improving the original task success rate. These findings highlight the promise of simple, modular, multi-layered defenses for strengthening the security and robustness of tool-integrated LLM agents. The code is available at https://github.com/Xiaoyan-Lisa/Defenses-for-Tool-Integrated-LLM-Agents-Against-Adversarial-Attacks.
Read the full abstract
Cosmology and Nongalactic Astrophysics
$\Lambda$CDM in a courtroom
The cosmological constant $\Lambda$, and its generalization as Dark Energy, makes up roughly 70% of the present-epoch universe and has been part of the dominant $\Lambda$CDM cosmological paradigm for more than a quarter of a century. It has been put on trial by recent measurements of galaxy clustering, Type Ia supernovae, and the cosmic microwave background. In this Perspective we evaluate the strength of the evidence for a constant $\Lambda$, versus Dynamical Dark Energy (DDE) in the light of recent developments. We also assess what the future holds both for our Universe and our understanding of its most mysterious component.
Read the full abstract
Software Engineering
Coaching Qwen3 Coder 30B to Think Like a CodeClash Arena Agent
Large language model coding agents have recently become useful for software tasks, but weaker or open-weight agents still struggle to reliably interpret user intent and execute complex multi-step workflows. This gap is especially visible in long-horizon settings, where an agent must repeatedly inspect prior outcomes, diagnose failure, and choose the next code edit under interaction constraints. It motivates a natural question: what can we do to improve the thinking process of a weak code agent? We study this question in CodeClash, a code-arena benchmark where the original work evaluates 8 commercial coding agents across 6 arenas through multi-round tournaments. Since Qwen3 Coder Plus ranks last among them, we take the open-weight Qwen3-Coder-30B as a case study and investigate how to improve it with distilled knowledge from stronger agents. Our analysis shows that Qwen3-Coder-30B is not well optimized for arena-style interaction: it frequently produces syntax and protocol-breaking errors and exhibits weak strategic adaptation across rounds. These failures are difficult to correct with vanilla instruction tuning alone, since offline SFT cannot directly verify whether a generated action is valid or beneficial. To address this, we propose ReAct SFT, which rewrites teacher trajectories into explicit [obs][thought][act] chains, and trajectoryquality weighted SFT, which reweights samples to encourage post-edit checking. ReAct SFT substantially improves strategic behavior, and our fine-tuned model outperforms the original Qwen3 Coder Plus in tournament evaluation.
Read the full abstract
Computation and Language
RAG-CT: Mitigating Privacy Risks on Retrieval-Augmented Generation Systems via Scanning Prompt Distribution
Retrieval-Augmented Generation (RAG) has emerged as a powerful paradigm for improving the quality of generated contents of Large Language Models (LLMs) by grounding responses in external knowledge, thus reducing hallucinations and factual errors. However, recent studies have highlighted a critical vulnerability: adversaries can exploit the retrieval process to extract personally identifiable information (PII) from the underlying corpus. To mitigate this risk, we propose a novel defense, RAG-CT, that identifies malicious queries by analyzing their entropy and margin distributions and using a score-based detection method. Extensive experiments with four state-of-the-art attack strategies and four defense baselines on two datasets show that our approach significantly reduces PII leakage while outperforming existing defenses. This work provides a lightweight yet effective mechanism to protect RAG systems against PII leakage without requiring modifications to the underlying LLM or retriever.
Read the full abstract
General Mathematics
Multiorder Fractional Operators: The Conformable Multiorder Derivative and Its Associated Multiorder Integral
This paper introduces a novel class of non-linear multiorder fractional differential operators. First, we provide intuitive and formal justifications for the existence of multiorder differential operators through differential analysis and Omega-derivative formulations. We then define the conformable multiorder derivative as a functional quotient of fractional derivatives and establish its analytical properties rigorously. Furthermore, leveraging fundamental principles of mathematical analysis, we construct its inverse operator, termed the conformable multiorder integral. Detailed step-by-step proofs of all theoretical properties are presented. Illustrative examples and applications to solving elementary multiorder differential equations are thoroughly examined. Finally, comprehensive conclusions and strategic outlooks for prospective research directions are discussed.
Read the full abstract
Machine Learning
Evaluating Open-Weight E-Commerce Agents with Environment-Grounded Verification
A shopping conversation has many routes to the same cart, and a task-success rate reduces all of them to one score. We build a deterministic and reproducible e-commerce environment that precommits each trial's customer and trajectory parameters, including the persona, difficulty, target cart, and an item reveal schedule. A simulated consumer attempts to buy a target cart from the environment with assistance from the evaluated model. The environment guides the simulator's actions and records every assistant action alongside the environment state at that point. After the trial, these records allow the evaluator to assess individual parts of the conversation against the retained evidence. For example, the evaluator penalizes a search for failing to surface a target product only when the customer has already mentioned that product. We further use this evidence to apply different penalties to tool calls depending on how the assistant's actions compare with an expected tool-call set. Our environment also interacts with the simulator bidirectionally, reading its output to stop the trial when the simulator determines that the customer has become too frustrated and injecting directives in real time that specify when to explore, defer buying an item, or recall a previous exchange. This interaction creates an open-ended and verifiable simulation. Across eight open-weight agents from 20B to 35B parameters, with 160 trials per agent and 44 metrics, the resulting capability profiles distinguish under-action, over-purchase, unsupported product attributes, and poor search, all of which terminal success obscures.
Read the full abstract
General Mathematics
From Umbral Hyperbolic Integrals to a Cotangent Coefficient Formula for the Mittag Leffler Polynomials
Let \(g_n(x)\) be the Mittag–Leffler polynomials defined by $$ \sum_{n\ge0}g_n(x)t^n=\frac12\left(\frac{1+t}{1-t}\right)^x. $$ We derive the coefficient formula $$ g_n(x)=\frac1n\sum_{j=0}^{\lfloor (n-1)/2\rfloor} (-1)^j\frac{2^{n-2j-1}}{(n-2j-1)!} [u^{2j}](u\cot u)^n\,x^{n-2j}, $$ equivalently $$ [u^{n-r}](u\cot u)^n =(-1)^{(n-r)/2}\frac{n(r-1)!}{2^{r-1}}[x^r]g_n(x). $$ The identity is suggested by comparing an umbral representation of the hyperbolic tangent integral $$ K(a,b)=\int_0^\infty \frac{\tanh^a x}{x^b}\,dx $$ with a compact cotangent-coefficient formula, but is proved independently from the generating function by formal Lagrange–Bürmann inversion. Substitution of the proved bridge back into the integral formula then yields a rigorous derivation of the umbral Mittag–Leffler representation. We also develop a Dirichlet–beta analogue for $$ I_{m,n}^{\beta}=\int_0^\infty\frac{\tanh^m x}{x^n\cosh x}\,dx. $$ Defining \(q_m(x)\) by $$ \sum_{m\ge0}q_m(x)t^m=(1-t^2)^{-1/2}\exp(2x\operatorname{artanh}t), $$ we prove $$ [u^{m-r}]\frac{u}{\sin u}(u\cot u)^m =(-1)^{(m-r)/2}\frac{r!}{2^r}[x^r]q_m(x). $$ This gives a parallel umbral representation for the beta integrals. The family \(q_m\) is a shifted \(c=-1,\beta=1\) specialization of the classical Meixner family, yielding a beta/Meixner counterpart to the zeta/Mittag–Leffler correspondence.
Read the full abstract
Machine Learning
Distilling Foundation Models for Agentic What-If Reasoning:Cost, Latency, and Governance in a Hybrid LLM+SLM Architecture
Tabular foundation models deliver strong zero-training predictive performance via in-context learning, but their high inference latency makes them impractical as hot-path decision backends in interactive agentic loops. We distill a TabPFN teacher into a compact feed-forward student across a business-decision simulation on UCI Adult and five OpenML benchmarks: the classification head compresses 53.2M parameters to 8,546 (6,220x); the deployed two-head loan pipeline compresses 111.4M parameters to 17,059 (6,532x). The student retains 95.4-100.5% accuracy and 96.8-100.0% AUC, with the lowest accuracy retention on credit-g at 95.4%; an alpha = 0 hard-label control shows that the teacher's soft targets provide a 2.1-7.0 AUC point gain.
Read the full abstract
Sound
MUUNRiver-Bench: Diagnosing Relation-Dependent Music Retrieval with Multimodal Instructions
Music retrieval is relation-dependent: given a reference track, a listener may seek its style with a new theme, a cover, or a comparable voice, and these intents demand contradictory rankings. We present MUUNRiver-Bench, a diagnostic benchmark whose reference-audio queries use natural-language instructions to define relevance. A pipeline combining expert genre priors, LLM-generated prompts and lyrics, synthesis, and expert review yields 3,440 tracks spanning 13 genres and 116 sub-genres, and seven tasks: similar-music, style-preserving lyric-rewriting, lyric-preserving style-rewriting, cover, vocal-timbre, isolated-vocal, and segment retrieval. Across six models in eight configurations, task-wise rank reversals reveal complementary biases: acoustic encoders favour local identity, whereas text-aligned encoders favour semantic relations. Frozen encoders diagnose default similarity preferences; instruction-aware and audio-text fusion systems provide exploratory tests of textual conditioning, with neither simple fusion scheme consistently improving its backbone
Read the full abstract
Robotics
Structure-Preserving Quantum Circuit Architectures for Robot Kinematics
Structured spatial data require quantum encodings that preserve geometric relations, expose measurable observables, and remain implementable on finite-depth hardware. This work introduces a quantum representation and circuit architecture for rigid-body transformations and specializes it to Denavit–Hartenberg kinematics of serial open-chain manipulators. Each translational contribution is factorized into a classical metric magnitude and a signed unit direction encoded by a single-qubit Bloch vector, while parameterized rotations reproduce the ordered propagation of frame directions. A selector register prepares probabilities proportional to the contribution magnitudes, and the reduced state of a designated readout qubit encodes their normalized weighted sum. The retained classical scale then reconstructs the metric end-effector position. Two additional readout qubits encode terminal-frame axes, providing a compact and geometrically interpretable pose interface. At the ideal expectation-value level, measured Pauli observables reproduce the corresponding classical kinematic quantities. Alternative circuit architectures realize the same representation with different tradeoffs in qubit count, circuit depth, controlled operations, and measurement requirements. Validation on a serial manipulator yields numerically negligible position and orientation reconstruction errors under ideal simulation. Finite-shot simulations, noisy executions, transpilation analysis, and a hardware demonstration further characterize statistical error, noise sensitivity, and implementation overhead without asserting computational advantage.
Read the full abstract
Social and Information Networks
netseg: a Python Package for Measuring Structural Polarization and Segregation in Social Networks
The study of structural polarization and segregation in social networks is an established line of research, and the quantification of both phenomena proceeds through a set of widely cited network indices. The code implementing those indices, however, is seldom released and almost never tested. We present netseg, a comprehensively documented Python package implementing these indices, most of them generalized to more than two groups and to directed as well as undirected input. It ports the R package of the same name and adds measures the R version lacks, among them Random Walk Controversy, Boundary Connectivity, Dipole Moment, and Moran's I. The package operates on igraph objects and performs the underlying graph operations (e.g., neighborhood queries and random-walk simulation) through igraph's Python interface, so that they execute in compiled code rather than in interpreted Python. For several of the indices this yields runtimes orders of magnitude below those of the available open-source implementations, which the documentation reports in benchmarks. Most of these indices are defined as a divergence from a null model, and published implementations fix that null model to a uniform random graph of matching density. netseg accepts an ensemble of graphs as a sample from an arbitrary null model, and distinguishes indices that already incorporate a baseline from those that do not, adjusting the comparison accordingly to avoid double subtraction. The documentation provides, for each index, a worked empirical example, its behaviour at the degenerate cases where it is undefined, a benchmark, and the procedure for substituting a custom null model. We report the behaviour of every index over a parameter sweep of a generative opinion model, and apply them to a county-level railroad network built from nineteenth-century operator records joined to full-count census data.
Read the full abstract
Physics and Society
How to build a campfire? Participatory modelling with justice
Energy transition decision-making is pervaded by deep uncertainties. Computational models are helpful in addressing such uncertainties, as they can give insight into techno-economic complexity. However, models alone are insufficient, as part of the uncertainties involve justice dilemmas.
Read the full abstract
Econometrics
Tractable Estimation of the Money Pump Index
The Money Pump Index (MPI) of Echenique et al. (2011) measures the severity of consumer irrationality, but computing the exact mean and median MPI over all revealed preference cycles is NP-hard (Smeulders et al., 2013). Existing solutions rely on heuristic proxies, such as evaluating only shorter cycles or bounding the MPI. By framing revealed preferences as a directed graph, this paper projects choice violations onto fundamental cycle bases, which are minimal sets of linearly independent cycles that span the graph's entire cycle space. This yields computationally tractable estimators for the mean and median MPI that are asymptotically equivalent to the original MPI. Applying this methodology to the scanner dataset analyzed by Echenique et al. (2011) and Smeulders et al., (2013), the proposed estimators compute quickly and with negligible small-sample bias.
Read the full abstract
Hardware Architecture
Is INT8 Portable? A Cross-Platform Measurement Study of Quantized Inference on Embedded and Automotive Accelerators
Eight-bit integer (INT8) post-training quantization is the default recipe for edge deployment, under a widely held assumption: INT8 makes inference faster at a small, predictable accuracy cost, and a model quantized once can be carried to any target. We test that assumption with a controlled measurement study across seven hardware classes – ARM and x86 CPUs, a discrete GPU, an NVIDIA Jetson AGX Orin iGPU and its NVDLA cores, and two vendor NPUs (Qualcomm Hexagon HTP, DEEPX DX-M1) – holding the ONNX artifact and the quantization scales fixed so the integer kernel or ISA is the only free variable. Portability fails on three axes. (1) The sign of the INT8 speedup is set by the CPU's dot-product ISA (ARM dotprod/SDOT, x86 VNNI): cores that have it speed up by up to 2.1x, cores that lack it slow down by 1.7x, for the identical model and runtime. (2) INT8 outputs are not portable, and the rule is an invariance rather than a gradient: FP32 predictions are bit-identical for every pair (1000/1000), while INT8 predictions agree 1000/1000 exactly when two targets share an integer kernel and 958-965/1000 whenever they do not – independent of whether the boundary is CPU<->CPU or CPU<->accelerator, and invisible to top-1 accuracy, which is preserved. (3) Vendor NPUs own quantization: a bring-your-own QDQ graph fails silently on one NPU (external scales ignored, accuracy 0.75 -> 0.005 while it compiles, profiles and runs without error) and loudly on the other (the compiler refuses the graph), so only the vendor's native path yields a correct engine. We further show that edge-NPU latency regimes are set by output/device-to-host transfer size rather than compute, and locate the transition with a fixed-compute sweep. We release the scripts and 32 reports. "Quantize once, deploy anywhere" is unsafe for embedded and automotive deployment, where per-input determinism and redundancy matter.
Read the full abstract
Superconductivity
Topological superconductors and Majorana fermions based on $X$-wave magnets with $X=p,d,f,g,i $
We study superconductors coupled with $X$-wave magnets with $X=p,d,f,g,i$. They are essentially different between altermagnets with $X=d,g,i$ and odd-parity magnets $X=p,f$ due to the compatibility of the spin-singlet formation of the $s$-wave superconductor. The altermagnets and odd-parity magnets show rich superconducting phase diagrams depending on the strength of the magnetic order, the chemical potential and the ratio of the on-site and nearest-neighbor interactions. Especially, we find that the altermagnetic order stabilizes the chiral ($p_{x}+ip_{y}$)-wave superconductor, which is a class D topological superconductor hosting Majorana chiral edge states. It is characterized by the Chern number. Especially, there emerge two chiral edge modes in the $i$-wave altermagnet, which are characterized by the Chern number 2. On the other hand, the $p$ ($f$)-wave odd-parity magnetic order stabilizes the $p$ ($f$)-wave topological superconductor, which is a class DIII topological superconductor hosting Majorana flat bands. It is characterized by the winding number. Majorana flat bands are robust in the presence of the mixing with the $s$-wave superconductivity, which is inevitable for odd-parity magnets although the magnitude of the mixing is tiny.
Read the full abstract
Combinatorics
Logarithmic Circumference In Tough Graphs
For every real $t>0$, we prove that every $2$-connected $t$-tough graph contains a cycle of length at least $\ell$ whenever $n \leq \ell(1 + 1/t)^{\lfloor \ell/2\rfloor - 1}$. This establishes the logarithmic bound conjectured by Broersma, van den Heuvel, Jung, and Veldman.
Social and Information Networks
Predicting Social Media Engagement using Machine Learning
Social media platforms are popular channels for disseminating information, owing to their large user bases and ease of access. Companies also use social media as an important aspect of the advertising process. By creating high-quality posts, companies can strengthen their engagement metrics and increase their follower count. While a growing body of research has examined social media engagement, fewer studies have jointly examined the visual, textual, and temporal features of image posts, even though these features collectively determine the performance of content on social media. To understand the important drivers of social media engagement, we collect image posts of furniture firms on Facebook and extract visual, temporal, and textual features from them using text and image analytics methods. We evaluate several machine learning models–including Random Forest, Light Gradient Boosting Machine (LightGBM), and eXtreme Gradient Boosting (XGBoost)–to assess the drivers and the prediction power of social media engagement using the features from our data. Our primary model, LightGBM, explains 81.8% of the variance in the engagement score, 2.3 times that of our baseline model. We observed that brand-relative deviation features, which capture how far a post departs from its brand's norms, are the strongest predictors, accounting for 51.3% of total importance. Our research quantifies the extent to which these features are associated with interactions and provides recommendations that organizations may consider.
Read the full abstract
Artificial Intelligence
Corrupt Plans, Clean Traces: Evading Chain-of-Thought Monitoring with Plan Injection
Chain-of-thought (CoT) monitoring is a safety strategy where the reasoning of a large language model "actor" is inspected by a "monitor" (often another language model) for signs of unsafe planning, deception, or misalignment. We find that planting harmful but benign-sounding reasoning in the actor's context can steer it to perform adversarial actions while evading monitors, an attack we term "plan injection". We initially discover this attack in the multiple-choice question-answering monitorability setting proposed by Lanham et al. (2023), using the investigator-agent elicitation framework of Li et al. (2025). We generalize the attack and show that the discovered behavior scales to harder tasks (achieving 25-33% monitor evasion rates across different monitorability benchmarks) and larger models such as DeepSeek-R1. Across the settings we study, actor models not only follow injected plans but also paraphrase them as their own reasoning, without explicit attribution to the injections. Finally, we find cases where extra monitor resources cause harm - giving the monitor access to the injected plan drops detection by as much as 50% in the Bio-Math task and in a case study on monitor reasoning budget, we find transcripts where additional thinking tokens are spent rationalizing the injected plan rather than flagging it.
Read the full abstract
Robotics
ResSafe: Learning Safety Filtering with Residual Reinforcement Learning for Humanoids
Safe control of humanoid robots remains challenging due to their high-dimensional dynamics, contact-rich interactions, and sensitivity to disturbances. Although reinforcement learning has enabled effective locomotion and motion tracking, learned policies can still generate unsafe actions that lead to instability or falls. In this work, we propose residual reinforcement learning as an implicit safety-filtering mechanism for safe humanoid control. Instead of relying on a single nominal policy to simultaneously balance performance, safety, and robustness, we decouple performance and safety. The nominal policy focuses solely on task performance, while a residual policy learns safety corrections. This decoupling leads to a better performance–safety Pareto trade-off and avoids the need for careful tuning of multiple competing reward terms within a single policy training. We show that the residual policy can act as an implicit safety filter.
Read the full abstract
Machine Learning
Bellman Policy Optimization
Reinforcement learning with verifiable rewards (RLVR) improves the reasoning capabilities of large language models (LLMs). We introduce Bellman Policy Optimization (BPO), a critic-free method derived from Policy Mirror Descent (PMD). For autoregressive generation with terminal rewards, BPO uses the Bellman equations to reformulate PMD as a trajectory-level objective. The reformulation avoids estimating state values at intermediate states. We prove that it has the same unique optimal solution as the original PMD objective. We derive the practical BPO loss by approximating this objective. Its mismatch-correction weight is a smoothed ratio of complementary token probabilities. Experiments on mathematical reasoning benchmarks demonstrate the effectiveness of BPO.
Read the full abstract
Quantum Algebra
Cyclic Haagerup-Izumi fusion categories at every odd order
We construct a complex spherical fusion category with cyclic Haagerup-Izumi fusion rules for every odd n >= 3, and deduce pseudo-unitary existence. The proof has four parts: explicit real coefficients and their quadratic identities; a contour calculation for a range of cubic Fourier coefficients; algebraic completion of all cubics; and categorical reconstruction. Matrix inversion supplies reflection, and an extension of the dimension-field automorphism supplies the positive-dimensional category.
Read the full abstract
High Energy Physics (Phenomenology)
Inferring dark matter masses and interactions from high recoil energy events in LUX-ZEPLIN
LUX-ZEPLIN has recently revealed a single nuclear recoil event compatible with a recoil energy of $E_R\simeq248\,\mathrm{keV}$, in a region with a small expected background. Interpreting this as a signature of dark matter, we compute the posterior probability distribution of the parameters of the dark matter scattering off xenon nuclei. We retain the exact two-body kinematics for an arbitrary mass splitting between the incoming and outgoing dark-sector states, thus encompassing both elastic and inelastic scatterings. The data select non-trivial, disconnected regions of this distribution, which correspond to physically distinct endothermic and exothermic regimes, in which the observed recoil is generated either close to a kinematic threshold or through the release of dark-sector mass energy. We determine the recoil spectra preferred by the observed data and show how the absence of accompanying events at lower recoil energies sharply constrains the viable parameter space. We find a kinematically tuned region of endothermic WIMP-like models at high mass for generic couplings, as well as exothermic models ranging from sub-GeV to heavy dark matter with sizable mass splittings. Elastic dark matter scatterings require instead momentum-suppressed and/or nuclear-spin-dependent non-relativistic interactions.
Read the full abstract
Probability
Convergence for Small-Order Derivatives of Random Polynomials with Independent Roots
Let $X_1,X_2,\ldots$ be i.i.d. complex-valued random variables with arbitrary Borel probability law $\mu$, and set $P_n(z)=\prod_{j=1}^n(z-X_j)$. For every deterministic sequence $k_n=o(n)$, we prove that the empirical zero measure of the $k_n$-th derivative $P_n^{(k_n)}$ converges weakly to $\mu$ almost surely.
Read the full abstract
Artificial Intelligence
Stellar Colosseum: A Many-Agent Harness for Long-Horizon Research in Mathematics and Theoretical Computer Science
Language models can produce plausible short proofs, but may still be unreliable on long-horizon research problems, where progress depends on a sequence of uncertain and interdependent decisions. We introduce Stellar Colosseum, a model-agnostic harness for allocating inference across research in mathematics and theoretical computer science. Colosseum explores alternative strategies before proof construction, uses a readiness gate to decide when a route is mature enough to decompose, represents the proof plan as interdependent section-level subproblems, and routes verifier findings back to the affected part of the argument. Across these stages, it generates candidates in parallel, attacks them with targeted falsification, and combines candidates and their critiques into a single research artifact through overlapping random-sample tree aggregation. The Colosseum workflow has been integrated into Google Antigravity's Teamwork framework as the Long Proof pattern. We demonstrate the capabilities of Colosseum through open-ended research and evaluations on theorem-proving and competitive programming benchmarks. Using Colosseum with Gemini 3.1 Pro, we obtain several new results that address open problems arising from papers published at top venues such as FOCS and JMLR. On TCS-Bench, a benchmark of research-level theorem-proving tasks drawn from papers published at FOCS, STOC, and SODA, Colosseum achieves 71.0% accuracy using Gemini 3.1 Pro and Gemini 3.7 Flash. In a separate Codeforces evaluation using Gemini 3.1 Pro, the proof-oriented pipeline with execution feedback solves 218 of 222 problems.
Read the full abstract
Machine Learning
The Router Within: Eliciting Native Skill Routing from a Frozen LLM
Skills extend an LLM agent beyond its parametric knowledge, and the gain they promise rests on picking the right one. Deployed harnesses route by preloading every skill's metadata into the context, which disperses the agent's attention and caps the library size. Retrieval pipelines move the selection out of the context, but also out of the agent's capability. We show that the frozen agent LLM already carries the routing signal in its own forward passes, and that two linear maps suffice to read it out with no skill text in the context. Our Gavel (Glance And Verdict from a frozen LLM) reads it in two steps. A glance scores the full library by matching the task's mid-layer states against a compact bank that one forward pass builds for each skill at installation, with the two maps as the only trained parameters. A verdict then resumes each shortlisted skill's forward pass, reads the model's own likelihood and yes/no judgment, and fuses both with the glance as a product of experts. Trained once, Gavel transfers zero-shot to three public benchmarks and SkillTraj, our new benchmark of 372 simulated agent trajectories. On Qwen3-32B it outperforms progressive disclosure and retrieve-and-rerank pipelines that add 1.2B to 16B external parameters, by up to 13.4 points on written tasks and up to 21.9 when the need for a skill arises mid-rollout. Routing accuracy improves as the backbone does, and in a bash-agent harness Gavel lets the 32B trigger the right skill on Skill-Use more often than models of up to 1.6T parameters in Codex.
Read the full abstract
Audio and Speech Processing
Acoustic Image Source Interpolation with Optimal Transport Barycenter
Room impulse responses can be estimated via the image source model (ISM) using the image source point cloud (ISPC) of a physical source. However, because the source movement changes the ISPC, estimating the ISPC at a new source position typically requires repeated acoustic measurements. We propose an optimal transport (OT) barycenter framework to interpolate the ISPC of a new source location from ISPCs of known sources. The method jointly estimates image-source associations and the ISPC at the new location. The OT ground cost exploits the property that the image sources undergo the same displacement as their physical sources. This approach is realized for both grid-based and support-free configurations. The support-free method addresses the resulting nonconvex joint estimation problem by alternating between identifying image-source associations across the ISPCs and refining the target image-source locations. This enables the interpolation of ISPCs without repeated measurements, facilitating efficient and flexible room-acoustic modeling.
Read the full abstract
Machine Learning
A Chosen Future Can Still Be Rewritten: Causal Writability in Video Models
When a video model generates physically incorrect motion, did it fail to learn the correct motion, or did it learn it but fail to use it? We show the latter: the correct motion remains available inside the model and can still be made to control the generated video. We train on videos where red masses oscillate slowly and blue masses oscillate quickly, then test a red mass with fast observed motion. Even when the model generates slow motion in this conflicting case, a low-dimensional edit predicted from simple physical variables restores the correct fast motion. We call this ability causal writability. At fixed strength, we find a sharp depth boundary: the same edit changes the video before the boundary but not after it. This closure marks commitment for that write. The motion signal nevertheless remains, and a stronger downstream write can restore physical motion, while excessive gain overshoots. Early causal writability predicts which errors training later corrects: those errors are writable at more network depths than errors that persist. We reproduce both causal writability and its sharp closure in a pretrained 1.3B video model, supporting generality across model scale and training regime.
Read the full abstract
Data Structures and Algorithms
The $k$-server conjecture is true
The $k$-server conjecture states that a deterministic online algorithm can achieve competitive ratio $k$ on every metric space. We prove the conjecture. Specifically, we show that the work function algorithm satisfies it. Our proof uses a natural algebraic representation of the work function as a matrix, which encodes all feasible paths to reach a configuration. In this representation, the minimum and addition operations arising in the definition of optimal costs correspond to addition and multiplication of formal expressions, and each work function value corresponds to the determinant of $k$ columns of the matrix. A request arrival updates the representation via a change of basis and row replacement. The amortized analysis is based on a potential function defined in terms of a larger matrix whose coordinates are pairs of coordinates of the original matrix representation.