SEARCH · Searcharxiv

Search Searcharxiv

Search indexed arXiv papers on artificial intelligence, large language models, computer vision and robotics. Read source abstracts and follow links to arXiv.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 19 recordsLinked to original sources

arXiv2026

Protocol effects on feature-based hardware-Trojan detection across Trust-Hub families

Trust-Hub reuses host circuits: several files differ mainly in the inserted Trojan. When gates from sibling variants enter both training and test folds, a detector can benefit from host logic it has already seen. We measure that effect instead of proposing another classifier. The corpus contains 49,124 gates from 16 netlists grouped into five host families. We left the parser, 36 gate features, class weighting, model settings, threshold, and family-level aggregation unchanged and altered one choice: the test boundary. The three settings draw test gates from the pooled corpus, withhold a complete netlist, or withhold every variant of one host. The choice matters. Random forest records F1/AP of 0.914/0.978 with pooled gates, 0.636/0.851 with one netlist held out, and 0.460/0.577 with a host family held out. XGBoost falls from 0.946/0.976 to 0.464/0.544 across the same comparison. Logistic regression loses AP, although its fixed-threshold F1 is not monotonic. Each family shows the same pooled-to-family direction. Feature removal, repeated model and simulator seeds, score normalization, parser-related exclusions, and a smaller sample change the size of the gap without reversing it. Aggregation also matters: a gate-weighted average is dominated by the larger ISCAS files, so the headline values give each host family one vote. Bootstrap and jackknife summaries keep the gap positive, but their folds reuse training families. We treat the five family rows as descriptive evidence rather than independent trials. Five host families are too few for a population claim, and the experiment says nothing about transfer to a new cell library or an industrial design. It supports a narrower conclusion: sibling benchmark variants can inflate apparent transfer. Benchmarks with several variants of one host circuit should report family-aware holdouts and all five family results beside pooled scores.

cs.LG↗

arXiv2026

Perturbation Sensitivity of Maximum-Likelihood Pairwise Ranking in Computational Decision Systems

Maximum-likelihood pairwise ranking is a com- mon computational mechanism for prioritization, reputation estimation, and comparison-driven decision support. Despite its broad use, the perturbation sensitivity of this estimator under structured changes in comparison data remains insufficiently characterized. We study this question as an applied-mathematics and computational-science problem in stability analysis. We for- mulate coordinated perturbation as a budgeted subset-selection problem over pairwise observations and introduce an Adaptive Subset Selection Attack (ASSA) as a scalable search heuristic for probing high-impact perturbation sets. Through experiments on synthetic and observed preference datasets, we show that MLE-based ranking can exhibit pronounced regime-dependent sensitivity: relatively small but coordinated perturbations may in- duce meaningful changes in output orderings, while the response profile varies across budgets and data conditions. By comparing ASSA with random, greedy, and randomized subset baselines under repeated trials, we characterize both the magnitude and the variability of perturbation-induced ranking shifts. These results position pairwise ranking sensitivity as a problem in computational reliability, numerical stability, and robustness auditing for engineering systems built on comparison-driven inference.

cs.LG↗

arXiv2026

Robust PAC Learning of Concurrent Stochastic Games

We introduce the first Probably Approximately Correct (PAC) learning framework for general-sum concurrent stochastic games (CSGs) with transition uncertainty, while addressing the challenge of Nash equilibrium (NE) existence. Our algorithm maintains data-driven $L^1$ confidence sets over transition kernels and solves a robust CSG to compute a social-welfare optimal $\varepsilon$-NE, using a robust MDP-based exploration mechanism to drive joint state-action coverage. Crucially, we introduce a Nash margin characterisation that enables principled reasoning about equilibrium existence: the framework either returns an $\varepsilon$-approximate NE whose social-welfare value is $\varepsilon$-close to optimal, or provides a sound certificate that no exact NE exists. Under a minimum reachability condition $p_{\mathrm{reach}}>0$ over relevant state-action pairs, the algorithm terminates after a polynomial number of trajectory samples, with sample complexity $\widetilde{O}\left( {R_{\max}^2 H^4 |S|^2 |A| / (p_{\mathrm{reach}} \varepsilon^2)} \right)$. Empirical results on benchmark CSGs demonstrate near-optimal performance, correct handling of equilibrium (non-)existence, and sample complexity consistent with theory.

cs.LG↗

arXiv2026

The Endogeneity of Miscalibration: Impossibility and Escape in Scored Reporting

An agent's probability report is paid for twice: by a strictly proper scoring rule, and by an approval rule for the decision it triggers. In this classical decision-coupled setting, non-affine approval is known to defeat truthful reporting. We show the conflict is endogenous: when feasible, the welfare-maximizing approval rule is never affine. The distortion, however, is predictable and can be designed around. There is a reserve report at which pretending to be the marginal type costs exactly the approval prize. Approving at or above the reserve screens types perfectly under every strictly proper score, and the reserve does not depend on the type distribution. A Lipschitz rule with a single kink attains first-best exactly; under strict feasibility no continuously differentiable rule does. The binding constraint is steepness, not smoothness. First-best is attainable within a slope budget if and only if the budget is at least the critical slope: the steepest chord of the pretending cost up to the reserve. Below it the welfare loss is cubic in the shortfall. Where the pretending cost is convex up to the reserve, as for Brier, log and power scores, the critical slope is closed-form. The instances are AI-agent oversight and marketplace operation.

cs.GT↗

arXiv2026

PokaiTrainer: Scaling Equilibrium Search to Competitive Pokémon VGC

Decision-time equilibrium search carried poker to superhuman play, but it has so far relied on tractable subgames: a handful of actions per decision, chance confined to card deals, one player moving at a time. Competitive Pokémon in its official doubles format (VGC) breaks all three assumptions at once. Both players act simultaneously from joint menus in the hundreds, each joint action resolves to hundreds of stochastic outcomes, and the opponent's reserves and stat allocations are hidden. No prior Pokémon agent performs equilibrium search, and whether it scales to this regime was open; we show that it does, and report what it took. PokaiEngine, our Rust battle engine, enumerates a joint action's full weighted outcome distribution in one pass, at ${\sim}99\%$ parity with Pokémon Showdown and a fraction of the cost of sampling it. PokaiTrainer adapts Student of Games to this scale and trains it by self-play over hundreds of human teams. Each decision is solved by counterfactual regret minimization as a Bayesian matrix game over public belief states, subgames grow under an explicit compute budget, and value targets are harvested from the interior of every solve and grounded by realized outcomes. The strength is in the search. The network's policy alone loses even to a shallow heuristic search. PokaiTrainer is, to our knowledge, the first VGC agent rated on the live Showdown ladder. Under open team sheets it wins 59% of 150 best-of-three sets against a human field averaging ${\sim}1320$ Elo, holds a 1350-1400 Elo band, and at its peak reached 1492 Elo, entering the format's top 500.

cs.LG↗

arXiv2026

Ascending Auctions for Combinatorial Markets with Frictions: A Unified Framework via Discrete Convex Analysis

We develop a unified ascending-auction framework for computing Walrasian equilibria in combinatorial markets with strong substitutes valuations and piecewise-linear payment functions. Our auction extends the celebrated ascending auctions of Gul and Stacchetti (2000) and Ausubel (2006) to accommodate payment frictions (e.g., transaction taxes or commission fees). This is achieved by incorporating directional price updates that reflect heterogeneous payment structures. Our framework also generalizes the unit-demand imperfectly transferable utility models of Alkan (1989, 1992) to a fully combinatorial setting, thereby unifying these paradigms. Furthermore, this is the first study to compute the minimum -- also known as the buyer-optimal -- equilibrium in combinatorial markets with such frictions. Our analysis builds upon discrete convex analysis. Our main technical contribution is a characterization of valid price-update directions, together with a strongly polynomial-time algorithm for computing them. Notably, the algorithm uses only demand- and exchange-oracle queries and never requires handling information of exponential size. To compute such a direction, we formulate a lexicographic extension of the polymatroid sum problem and characterize its dual solution via a reduction to a convex flow problem. Exploiting the $\text{L}^\natural$-convexity of the dual objective, we show that the desired direction can be constructed from the minimal dual solution. This convexity also yields transparent economic and potential-based interpretations of the auction dynamics, strengthening the connection between ascending auctions and discrete optimization.

cs.GT↗

arXiv2026

Exact-Form Regret for Gradient Descent, Mirror Descent and Follow-the-Regularized-Leader

Online gradient descent is usually studied through external regret, where the learner competes with fixed alternatives. Recent work shows that first-order methods control richer action-dependent deviations. We ask for a geometric characterization of the deviations with respect to which online gradient descent, mirror descent, and follow-the-regularized-leader (FTRL) achieve no regret. We identify exactness as the common principle. Exactness means that the relevant displacement field is generated by a scalar potential, or equivalently that the associated one-form is exact in the geometry used by the algorithm. This geometry depends on the algorithm. For gradient descent it is Euclidean geometry, for mirror descent it is the geometry induced by the regularizer, and for FTRL it is the cumulative dual state. Under mild regularity conditions, exactness yields sublinear regret, while nonzero circulation provides the complementary obstruction and leads to linear regret. This gives a unified geometric framework for understanding the deviation classes controlled by these algorithms and reveals that different first-order methods can control genuinely different classes of deviations. These deviation classes have direct consequences for learning, particularly in games. We study the equilibrium notions induced by exact-form deviations and introduce conservative correlated equilibrium, reflecting both the conservative geometry of the underlying displacement fields and the restricted family of deviations available to the players. We characterize its relation to correlated equilibrium, determine when the resulting equilibrium notions coincide and when they separate, and show how these relationships depend on the geometry and the learning algorithm. Overall, this work gives a unified geometric account of what first-order online learning algorithms are no-regret with respect to, beyond fixed comparators.

cs.LG↗

arXiv2026

Looking for Bidding Teammates: A Game-Theoretic Model of Stranger Collusion in Peer Review

Paper bidding is the entry point to reviewer assignment at large CS conferences: reviewers declare interest, combined by an optimizer with automated affinity scores. Reviewers who have never met recruit each other online, exchange identifiers, bid on each other's papers, and reciprocate with inflated scores. Existing collusion models assume already-trusted colleagues; open recruitment removes that assumption and the mechanism that made such arrangements work. We give the first game-theoretic model of collusion \emph{formation} in peer review, the \emph{Mutual Bidding Dilemma}: a four-stage game covering recruitment, exchange of identifiers under risk of being reported, unverifiable bidding, and reciprocal reviewing. The model predicts the arrangement cannot form: once assigned a partner's paper, writing the inflated review is pure cost, since the benefit depends on the partner's own decision. Reciprocation is never individually rational, for any payoffs, and the arrangement unwinds. What closes the gap is enforcement, not incentives: authors see their own reviews, deadlines recur every few months, and the group remembers who reciprocated. We derive the condition under which inflation is sustainable, the detection rate above which no partnership survives, and show effort enters both. On a calibrated end-to-end conference simulation, no detector we test exceeds $F_1 = 0.322$ against an attacker who camouflages bids, spreads them around a ring, and manipulates affinity; a two-person arrangement is worth $3.2$ points of acceptance probability; and the harm is distributional, not aggregate: $70$ honest papers are displaced while mean quality moves by only $0.002$, so no summary statistic reveals it. Randomized assignment is the one defense reaching enforcement itself, making a partner who never bid indistinguishable from one who bid and lost.

cs.GT↗

arXiv2026

Input-to-State Stability Framework for Fully Distributed Primal-Dual Dynamics for Quadratic GNEPs Without Multiplier Consensus

Generalized Nash Equilibrium Problems (GNEPs) often arise in multi-agent engineering applications that require distributed algorithms. Unlike traditional approaches that enforce consensus on multipliers, our method removes the need to share multipliers, reducing communication and improving privacy. As a result, different initializations can lead to different GNEs, including non-variational ones. We establish convergence under sufficient conditions using an input-to-state stability (ISS) framework.

math.OC↗

arXiv2026

Adaptive Strategies for GR(1) Games

We consider two-player GR(1) games on graphs, where the system player Eve must satisfy \[ \Box\Diamond A_1\land\cdots\land\Box\Diamond A_m \;\implies\; \Box\Diamond G_1\land\cdots\land\Box\Diamond G_n \] against the environment player Adam. Here $A_1,\ldots,A_m$ are assumptions on the environment, $G_1,\ldots,G_n$ are guarantees the system must provide, and $\Box\Diamond S$ denotes ``always eventually $S$''. Traditional static strategies are overly conservative: they may actively violate assumptions to trivially satisfy the implication, or abandon all guarantees when any assumption is violated. Existing methods to prevent such behaviors incur doubly exponential blowup. We introduce an adaptive framework treating Adam as a non-adversarial agent with unknown objectives. Eve monitors which assumptions Adam actually meets and adapts her strategy at runtime to maximize satisfied guarantees. Central to our approach is a novel algorithm for monitoring liveness properties $\Box\Diamond S$, enabling Eve to maintain real-time likelihood estimates of which assumptions will be fulfilled. Eve pre-computes strategies optimal for different assumption subsets, deploying a probability distribution over them that dynamically adjusts based on monitor outputs. We prove that when assumptions are violated, Eve's randomized adaptive strategy converges asymptotically to the deterministic strategy maximizing guarantees. A prototype demonstrates effectiveness and superior computational performance compared to the state of the art.

cs.LO↗

arXiv2026

The Non-Orientable Topology of Condorcet's Paradox

Preference cycles are prevalent in problems of decision-making, and are contradictory when preferences are assumed to be transitive. This contradiction underlies Condorcet's Paradox, a pioneering result of social choice theory, wherein intuitive and seemingly desirable constraints on decision-making necessarily lead to contradictory preference cycles. Topological methods have since broadened social choice theory and elucidated existing results. However, characterisations of preference cycles in topological social choice theory are lacking. In this paper, we address this gap by introducing a framework for topologically modelling preference cycles that generalises Baryshnikov's existing topological model of strict, ordinal preferences on 3 alternatives. In our framework, the contradiction underlying Condorcet's Paradox topologically corresponds to the non-orientability of a surface homeomorphic to either the Klein bottle or real projective plane, depending on how preference cycles are represented. These findings allow us to reformulate Arrow's Impossibility Theorem in terms of the orientability of a surface as well.

math.AT↗

arXiv2026

The Internal Anatomy of Strategic Choice in Large Language Models

Large language models act as strategic agents and models of human choice, yet choosing like a strategic agent does not mean computing like one. We recorded activations from four open-weight models --- dense and mixture-of-experts, including a matched base--instruct pair --- in one-shot play of 144 strict ordinal $2\times2$ games. We followed a prespecified incentive from prompt, through activations, to choice. Dense models mirrored the unadjusted human decline with game complexity. Incentive and choice were detectable in every model, but models differed in whether incentive reached the choice, aligned with it and, where tested, whether strengthening it shifted preference. The base and instruction-tuned Qwen2.5 models chose almost identically at baseline yet differed in whether incentive reached choice. Fixed decision cues were distinguishable internally but changed choices selectively. Similar behaviour can rest on different computation; post-training can reshape the path from represented incentive to decision while leaving behaviour and decodable information largely intact.

cs.AI↗

arXiv2026

Networked Multi-Resource Defense Capabilities in a General Lotto Game

Ensuring the security of complex systems involves the strategic allocation of defensive resources to prevent various types of attacks from succeeding. A defender often has multiple types of defensive assets at its disposal, where it must decide how to optimally deploy their heterogeneous capabilities across different attack types. In this paper, we formulate a multi-resource allocation problem in the form of a General Lotto game where a defender possesses various types of resources. A feature that we introduce is that their individual effectiveness against different types of attacks is characterized by a network weight matrix. In our analysis, we derive upper and lower bounds on the performance of the defender, and provide numerical evidence suggesting that they are tight. For the case of two attack types, we analytically prove that the bounds coincide, establishing an exact equilibrium characterization. We then numerically compare our proposed networked multi-resource architecture to an independent-defense benchmark from the existing literature. These results highlight fundamental and tractable structures underlying multi-attack-type defense problems.

cs.GT↗

arXiv2026

Closing Gaps in Online Fair Division

We study the online fair division of indivisible items, where items arrive one at a time and must be allocated immediately and irrevocably. We address three central open questions in the literature. First, we show that for every $n\ge 2$ agents, every fixed $k\ge 1$, and every $α\in(0,1]$, no online algorithm can guarantee $α$-PROP$k$ against an adaptive adversary. This remains true even when the total number of goods is known in advance, all values lie in $[0,1]$, and every good is positively valued by at most two agents. The impossibility extends to a broad range of standard envy-based, proportionality-based, and share-based fairness notions. We also establish an analogous impossibility for chores. Second, in the setting with predictions, a lightweight form of future information, the maximum item value, was previously known to guarantee only $1/n$-PROP1, leaving open whether the dependence on $n$ is necessary. We give a deterministic $19/30$-PROP1 algorithm against adaptive adversaries that does not require knowing the total number of goods. Given an additional upper bound $κ\in[2,n]$ on the number of agents who value any good positively, the guarantee improves to $\max \{ 19/30, n/(n+κ) \}$. The guarantee remains a positive constant under any fixed one-sided prediction error below one. When predictions are exact and the total number of goods $m\ge n\log n$ is known in advance, a deterministic algorithm achieves the same $19/30$-PROP1 factor together with $O(\sqrt{m\log n/n})$ maximum additive envy after normalizing each agent's values by their maximum item value. Third, against a non-adaptive adversary, we determine the tight high-probability PROP1 guarantee of the classical Like rule, which assigns each good uniformly among the agents who value it positively. Its guarantee improves when fewer agents value the same good, unlike uniform random allocation.

cs.GT↗

arXiv2026

Stable Voting Rules on the Edge of Optimal Metric Distortion

We prove the existence of a randomized voting rule with metric distortion at most $2.13713$, within $0.025$ of the lower bound of $2.11264$. Our rule comes from a generalization of stable $k$-lotteries developed in the context of committee selection. In contrast to prior work, our rule samples from a single distribution derived from a zero-sum game, without mixing between voting rules. Our result also gives sharp distortion bounds for stable $k$-lotteries, and in particular shows that stable $2$-lotteries have distortion $7/3$, despite only relying on aggregate preferences over triples of candidates.

cs.GT↗

arXiv2026

Subquadratic Subsidies for Nonnegative or Nonpositive Valuations

We study envy-freeness with subsidies for indivisible items beyond additive valuations. Assuming that every single-item marginal value lies in $[-1,1]$, we prove that a total subsidy of $O(n^{3/2}\sqrt{\log n})$ suffices to achieve envy-freeness among $n$ agents whenever all agents assign nonnegative values to every bundle or all assign nonpositive values to every bundle. These valuation classes include monotone goods and monotone chores, respectively, but do not require monotonicity. Our result establishes the first subquadratic total-subsidy bound for general monotone valuations that holds for every number of agents.

cs.GT↗

arXiv2026

Jointly Satisfying Pareto Optimality and Justified Representation is NP-Hard in Approval-Based Multiwinner Voting

An open problem in approval-based multiwinner voting concerns whether we can efficiently compute committees that satisfy both justified representation and Pareto optimality. We answer this question negatively by proving that, on the domain of all profiles, outputting a committee satisfying both axioms is NP-hard. An initial proof was found by ChatGPT Astra. This was then verified and rewritten by the author.

cs.GT↗

arXiv2026

Rank Without an Oracle: Deviation-Aware Interaction-Rank Selection from Offline Multi-Agent Logs

Offline multi-agent payoff models are estimated under a logging distribution but used on distributions induced by learned solutions and unilateral deviations. Standard held-out loss can therefore favor an interaction class that predicts logged play well while distorting strategic incentives. We introduce Selective Interaction-Rank Validation (SIRV) for finite games with known logging distributions. A training split fits nested payoff models and constructs a common union of all candidate deployment and unilateral-replacement distributions; an independent calibration split evaluates every candidate on this same union. SIRV returns the smallest rank whose simultaneous upper worst-target risk is within tolerance of the best upper score, and abstains when a declared target is unsupported or too imprecisely estimated. A common coverage event yields a finite-candidate target-risk bound and a candidate-specific coarse correlated equilibrium (CCE) gap certificate. We also isolate an exact two-point off-support non-identifiability result. In a controlled factorial study with 2,048 independent games per family, empirical-Bernstein bounds reduce the median CCE-gap certificate by 42.5% relative to Hoeffding bounds on common returns, with a 1.36-point reduction in supported return. Under paired rank misspecification and in a separately generated congestion family, the SIRV-EB fallback rule lowers mean true candidate-selection CCE regret relative to ID-Mean, while retaining game-level losses. Across 384 games at $N=3,5,8$, ID-Mean-relative mean CCE-regret effects stay positive while certified return falls sharply under weak coverage. These results separate certifiable model selection from universal strategic improvement.

cs.MA↗