Papers

In this paper, we extend earlier findings of supra-competitive outcomes in optimal-execution games by identifying a learned punitive mechanism that deters deviations and provides behavioural evidence of collusion. We investigate this mechanism in a two-player, finite-horizon Almgren-Chriss liquidation game. Independent proximal policy optimisation agents with access to within-episode price and action histories achieve costs below the Nash benchmark. We identify a profitable deviation by training against the mean learned liquidation schedule, then impose its first trade on one of the original agents. The opponent responds by accelerating liquidation. This response more than offsets the deviator's gain in every run and both player roles, while leaving the punisher's average payoff materially unchanged relative to not punishing under the same deviation. The punisher imposes greater losses on the deviator while preserving its own average payoff, despite the availability of more profitable, less punitive liquidation plans. Matching deviations and subsequent additional selling rise and later decline during training, while final policies retain an effective punitive response. We formalise two checks: whether punishment outweighs the gain from deviating, and whether the change in trading behaviour is large enough to account for the loss imposed. Both checks hold for the tested deviation. Together, these findings provide behavioural and economic evidence supporting a collusive interpretation of the learned supra-competitive outcomes.

Wave-based coherent imaging, including terahertz tomography, synthetic-aperture acoustics, and millimeter-wave radar, forms images by Fourier-processing finite-length signals, with an exact point spread function that is not Gaussian but a Dirichlet kernel: complex-valued, oscillatory, and periodic. However, transplanting 3D Gaussian splatting to coherent sensing fails by construction; Gaussian splats discard the sidelobe energy (10-20% of the total) and the phase that governs coherent interference between reflectors. Our key idea is to replace the learned Gaussian footprint with the physically exact Dirichlet kernel of the finite-window DFT, modulated by a surfel that carries area, normal, and material, so that the rendering primitive matches the measurement physics instead of approximating it. We pair this primitive with a specialized solver, Dirichlet Sliding Frank-Wolfe (DSFW), that combines variable projection, residual dual certificates, and certificate-driven hard replacement of low-utility surfels, with periodic low-resolution coupled Levenberg-Marquardt correction, navigating the rugged loss landscape that breaks generic first-order optimizers. The Dirichlet kernel admits an O(1) closed-form evaluation, so the forward model matches FFT ground truth to machine precision while remaining differentiable end-to-end. On dense terahertz reconstruction, our method recovers reflector centers to 0.018 bin RMSE, 10-50x faster than waveform-level automatic differentiation, where Gaussian splats fail.

One might think that learning the identity function with a deep linear residual network is trivial - the path along residual connections already implements the identity, and so the network need only drive its weights to zero. However, this zero-weight solution is just one point on an entire manifold of population-loss minimizers, each corresponding to a different decomposition of the identity across the network's layers. Although the population loss does not distinguish among these solutions, stochastic gradient descent (SGD) reproducibly favors particular ones. For instance, under anisotropic label noise, the learned layers exhibit a noise-dependent spectrum; even with weight decay, SGD does not generally recover the zero-weight solution. Changing only the parametrization, while leaving the set of realizable functions unchanged, yields different behavior: factoring each weight matrix as a product of two matrices causes the weights to collapse to zero, even without explicit weight decay. While perhaps mysterious and unintuitive at first, these phenomena can be understood through the lens of entropic loss, which augments the population loss with a term proportional to the expected squared norm of the minibatch gradient (Ziyin et al., 2025). On the identity manifold, the population loss is constant, while the entropic term distinguishes among these decompositions. We characterize its minimizers analytically and use them to derive predictions for the structure of solutions favored by SGD. Networks trained with SGD closely match these predictions. Overall, the identity learning task studied here serves as a clean and simple case study of how the lens of entropic loss can clarify why SGD favors particular decompositions of the same input-output function.

Large language model (LLM) based agents are often criticized for lacking spatial understanding and mainly exploiting statistical text patterns. We investigate their spatial comprehension through an architecture combining geometrical tools with a LLM serving as a high-level orchestrator in grid-world environments. The agent first collects geodesic trajectories, which are then vector-quantized to extract a representative subset. Offline, the LLM associates a natural language description of the underlying behavioral patterns to each selected trajectory, making it a tool. Online, the LLM chooses the appropriate tool conditioned on the current state and goal. Low-level control is handled by primitive actions that execute the trajectory associated with the tool. From an agentic AI perspective, this approach separates learning into two levels: tool discovery is handled through unsupervised quantization of trajectories, while reasoning and decision-making are handled by the LLM. We test the approach in a partially observable dynamic 2D grid environment with an open vision-language model (Qwen3.6-35B-A3B). Pairing the geometry-derived tool library with an agent-centered zoom tool and a collision detection tool lets a fast, non-reasoning configuration match the goal-reaching rate of a much more costly chain-of-thought version, while cutting the cost of a decision from minutes to seconds.

Explainable suicide-risk assessment requires models not only to estimate risk severity, but also to identify supporting language and the risk and protective factors expressed in a post. We present our system for the IEEE BigData 2026 Cup on Explainable Suicide Risk Assessment on Social Media, which addresses three tasks: risk-level classification, evidence phrase extraction, and multi-label factor identification. Our approach adapts Qwen2.5-Instruct models using quantized low-rank adaptation (QLoRA) and an answer-masked causal language-model objective. We jointly train across all three tasks for risk classification, jointly train on Tasks 1a and 1b for evidence extraction, and adapt Task 2 separately for factor identification. We also tailor aggregation to each output: we average risk-level probabilities from the 32B and 72B models, combine evidence phrases through cross-fold consensus, and calibrate factor-specific decisions through rate matching based on out-of-fold operating points. On the official leaderboard, the final system achieved a composite score of 0.7738, with 0.8089 on Task 1 and 0.6919 on Task 2. Across the evaluated configurations, three-task training performed best for Task 1a, joint training on Tasks 1a and 1b performed best for Task 1b, and task-specific training performed best for Task 2. Probability averaging further improved Task 1a when component models had complementary errors. These findings highlight the value of tailoring both training objectives and aggregation strategies to the output structure of each task within a unified language-model framework.

Legal research is a core and time-consuming legal workflow. Lawyers must identify controlling authority, verify that it remains valid, reconcile statutes and cases, and synthesize a grounded answer. Language model agents are a natural fit for this retrieval-intensive workflow, and automating even part of it would be valuable. But that value depends on reliability: a single missing authority, stale citation, or wrong legal conclusion can make an otherwise plausible answer unusable. We introduce Legal Research Bench (LRB), a benchmark of 413 open-ended U.S. legal research questions written by experts, each paired with a gold answer, supporting authorities, and a binary grading rubric. We evaluate thirteen frontier models in a harness with web search, case-law search, page parsing, and retrieval tools. We score agent responses through all-pass grading with source verification, where a response is correct only if every required criterion is satisfied and its cited authorities verify. We also validate the LLM judge against expert attorneys ensuring that benchmark scores track attorney judgment. Agents remain far from reliable: among the models we tested, the strongest, Claude Opus 4.8, is fully correct on 42.9% of questions. Performance also varies substantially by task setting: all-pass rates differ across areas of law and are lower on questions requiring reconciliation of conflicting authorities. Across models, more turns, tool calls, and inference cost do not predict higher accuracy.

Agents competing against a common opponent can often gain an edge by forming alliances, but deciding whether to do so is difficult when they lack precise knowledge about their opponent or ally. We study how uncertainty affects opportunities for alliance formation in the coalitional General Lotto game, in which two players compete aganist a shared adversary by allocating their resource budgets over separate sets of valued contests. When all agents know one another's budgets, it is known that a budget transfer from one player to the other can benefit both players. We ask whether such alliances survive when the players know the relevant budgets only up to a multiplicative factor. Under two models, one in which the players are uncertain about the adversary's budget and one in which they are uncertain about each other's budgets, we characterize exactly which games admit a transfer that benefits both players in every game consistent with what they conjecture to be possible. In both models, we show that such robust transfers exist in a set of games of positive measure for every level of uncertainty. These results suggest that alliances can remain worthwhile even when allies are largely in the dark.

Historical music restoration (HMR) has almost exclusively focused on constrained problems such as Super-Resolution or the restoration of solo pieces, under-exploring the general task of restoring orchestral historical music, which has multiple instruments. This under-exploration is largely because the HMR domain, early-20th-century recordings, has no pre-degradation ground-truth pairs, making the restoration task unsupervised and more challenging. This paper presents a supervised end-to-end orchestral HMR benchmark by exploring both the synthetic degradation functions and the end-to-end generative deep-learning restoration methods. We simulate the historical recording degradation chain more faithfully than prior work, which makes orchestral restoration into a tractable supervised problem. A latent flow-matching model trained on the resulting synthetic pairs outperforms existing HMR baselines on intrusive, non-intrusive, and subjective evaluations. We also curate and release a 9.3-hour license-free, unpaired, historical classical-music test set, along with code and audio demos.

Large Language Models increasingly serve as interfaces for knowledge-intensive information seeking tasks across languages by synthesizing multilingual evidence. Prior work has shown that they often exhibit query-language preference – the tendency to favor sources written in the language of the query – but has largely examined this behavior in settings where equivalent knowledge is available across languages. However, this bias becomes consequential when sources in different languages provide incomplete or inconsistent accounts of the same fact, since the information users receive then depends on the sources a model selects to use. To characterize query-language preference under such cross-lingual knowledge disparities, we introduce Waldo, a multilingual Question-Answering (QA) benchmark constructed from Wikipedia. Waldo contains 12K QA pairs targeting knowledge gaps, where a fact is available in one language but absent in another, and knowledge conflicts, where language editions provide conflicting versions of the same fact. Evaluating eight models across five languages, we find that when one language edition merely lacks the relevant fact, models generally use evidence from the other language regardless of the query language. Under conflicting accounts, however, model responses strongly align with the document in the query language, causing semantically equivalent queries to elicit different accounts depending on the user's language. Finally, we explore two different approaches that could mitigate this preference under knowledge conflicts: a mechanistic intervention that ablates attention heads associated with query-language preference, and LoRA-based training, which reduces the preference gap by up to 61.5%.

The rapid proliferation of data centers (DCs), driven by cloud computing and artificial intelligence (AI), has led to massive energy demand and carbon emissions, posing significant sustainability challenges. Carbon-aware optimization in geographically distributed data centers has been widely studied. Most existing approaches mainly focus on operational carbon emissions from server usage. However, existing literature often ignores workload-induced thermal stress, which accelerates nonlinear hardware degradation. This leads to more frequent server replacements and ultimately increases embodied carbon emissions. To address these limitations, we propose a comprehensive carbon life-cycle modeling framework for distributed data centers. Apart from operational carbon emissions, this work combines workload scheduling with a utilization-dependent exponential aging model to evaluate long-term carbon costs from server degradation. In order to solve the proposed optimization model in an online and privacy-preserving manner, an enhanced Lyapunov framework with time-varying queue shifting (TVQS) is first introduced to handle system uncertainties. Then, a zero-sum perturbation-based alternating direction method of multipliers (ZSP-ADMM) framework is developed to enable distributed coordination across geographically separated data centers while protecting locally exchanged workload information. Simulation results demonstrate that the proposed approach achieves up to 13.0% lower carbon emissions and 12.6% lower operational costs compared with benchmarks.

Vision-language-action policies often see only one or a few recent frames, which makes it difficult to evaluate how they use information that disappears during a task. We introduce MIKASA-Robo-VLA, a benchmark of 90 language-conditioned manipulation tasks. All but 10 hide the cue an action depends on. Those 10 are reactive controls. MIKASA-Robo, the suite it rebuilds, has 32 tasks and uses language only in a representative VLA subset. Here every task provides an instruction, while memory-dependent tasks hide a task-relevant cue and reactive controls keep it available. For 70 tasks, environment phase timings specify an information gap, and for 28 of them the gap exceeds the 16-frame window of the widest fixed-context VLA we survey. The gap counts only the interval the cue is provably absent, not the full duration a policy must retain it, so every memory-dependent task still requires memory by construction, including the ones whose measured gap is short. We release 22,500 oracle trajectories across 10 memory types in RLDS and LeRobotDataset v3. A reference $\pi_{0.5}$ baseline with current images and proprioception, but no observation history or explicit memory module, is fine-tuned on 14 tasks and achieves 0.211 $\pm$ 0.044 mean task success. Its lower success on the evaluated Long-split tasks is confounded by open-loop chunking and the memory types represented in that subset. Project page: https://mikasarobo.github.io/

Reasoning-enabled VLA policies expose chain-of-thought (CoT) traces that appear to explain and guide their actions, creating a potential interface for runtime safety through reasoning monitoring and correction. In this work, we define and operationalize two evaluation axes for assessing when this interface can improve embodied behavior: correctability, which measures whether unreliable reasoning can be detected and improved during generation, and actionability, which measures whether reasoning corrections produce behaviorally meaningful changes in the intended direction. To enable correctability, we introduce Token-level Reward for Utility-Steered Chain-of-Thought (TRUST), an offline-trained value model that predicts eventual reasoning correctness from partial prefixes and uses these estimates to monitor and selectively steer reasoning generation in frozen VLA policies. On the Alpamayo 1.5 driving VLA, TRUST monitors correctness with 88.9% accuracy and improves reasoning correctness from 75.9% to 90.0%. On a baseline-defined challenging subset in AlpaSim, TRUST reduces collision rate by 30.4% and maximum trajectory error by 11.5% relative to the unsteered policy, outperforming a compute-matched Best-of-4 baseline. On the DeepThinkVLA manipulation VLA, TRUST improves the correctness of grasp-state claims from 69.3% to 90.2% and action-choice claims from 68.8% to 85.9%, yet closed-loop task performance on LIBERO-Plus remains largely unchanged. Empirical analysis reveals intent-consistent behavioral effects in Alpamayo 1.5 but limited effects in DeepThinkVLA, helping interpret these different task-level outcomes. Together, our results show that gains in reasoning correctness do not automatically imply gains in embodied performance, motivating evaluation of correctability and actionability when using CoT as a runtime safety interface.

Representation alignment has become an effective way to accelerate diffusion training, but its benefits do not transfer reliably to pixel-space clean-image prediction. In JiT, we find that auxiliary feature alignment can improve access to semantic features while reducing access to image variation needed for clean-image prediction, creating a mismatch between the auxiliary objective and the denoising task. This suggests a different principle: auxiliary supervision should improve the prediction target itself rather than impose a separate representation target. We introduce JAx (Just Align x), a prediction-supervision method that aligns clean-image predictions across noise levels. JAx couples a noisier student observation with a cleaner observation through a Markov degradation that preserves the original JiT input distribution. Under this coupling, the oracle prediction from the cleaner state has the same conditional mean as the optimal JiT target, while its conditional target covariance is no greater. Thus, oracle prediction alignment preserves the population JiT objective up to a constant while providing a lower-variance training target. To make this construction practical with an imperfect EMA teacher, JAx combines ground-truth supervision with a reliability-gated coupling band that selects nearby teacher states based on prediction risk. On ImageNet 256x256, JAx consistently improves FID and accelerates convergence across JiT-B/16, L/16, and H/16, without an external encoder or changes to the architecture or sampling procedure. Gradient diagnostics further show reduced minibatch gradient variance, while ablations demonstrate that the gains cannot be explained by time reweighting alone. These results show that prediction-space supervision provides a simple and principled alternative to representation alignment for pixel-space generative models.

In frequency division duplexing (FDD) massive multiple-input multiple-output (MIMO) systems, the uplink (UL)-to-downlink (DL) channel covariance matrix (CCM) conversion problem is studied to relieve the heavy burden of DL training and feedback required for channel estimation. Learning- based methods perform well up to a certain array size, but for a fixed dataset size their accuracy deteriorates with the number of antennas, to the point where simple model-based methods outperform them. This paper identifies a key cause of this behavior and addresses it. The mean angle of arrival (AoA) induces a phase ramp along the lags of the CCM. Since the oscillation rate of this ramp grows with the number of antennas, a dataset of fixed size becomes increasingly sparse relative to the variation that must be captured. We propose estimating the slope of this ramp from the UL CCM separately and mapping it to the DL band in closed form, leaving the learner with a residual that is largely insensitive to the mean AoA, which substantially reduces the performance degradation with an increasing number of antennas. The proposed scheme, termed deramping, is a combination of pre- and post-processing steps that applies to learning-based conversion methods without altering their internal structure, as demonstrated on three structurally different learners. Simulation results show that deramping reduces the covariance estimation error of all three learners under uniform, Laplacian, and Gaussian angular power spectra,keeps the interpolation-based learners ahead of a model-based benchmark at large array sizes, and improves downlink channel estimation.

In partially observable reinforcement learning (RL), a later observation can make stored information obsolete or change what it implies for the next decision. Memory architectures and benchmarks for RL mostly test retention, the ability to keep information unchanged until it is needed. We formalize two further requirements. Rewriting sets the decision-relevant content to a value independent of the old one, and experience fusion transforms the old content by a rule that a later observation specifies. For tasks built from such updates, we count the memory states that a solution needs, and several baselines reach their lowest success rates on compositions that need more states. We introduce ALER (Adaptive Learnable Experience Rewriting), an agent that pairs an LSTM with a slot memory. An independently addressed Gumbel-Softmax write that concentrates its weight on one slot overwrites that slot, and a learned gate fuses the retrieved content with the recurrent state before the policy and value heads. We also introduce Rune-Mazes, three environments in which rune observations invert, cancel, reset, or repeat updates of a hidden cue under vector and pixel observations. Against seven baselines, ALER reaches a success rate of at least $0.82$ in all sixteen Endless T-Maze configurations and at least $0.99$ on all five Rune T-Maze compositions, and it has the highest mean success rate on four-branch Rune Multi-Corridor with an Invert rune. On pixel-based Rune MiniGrid Memory, it has a higher mean success rate than PPO-LSTM in eight of ten configurations. Project page: https://quartz-admirer.github.io/ALER-Adaptive-Learnable-Experience-Rewriting/.

An autonomous cyber defender trained with reinforcement learning (RL) is typically tied to the network on which it was trained, limiting its ability to generalize as network scale changes. Hierarchical RL reduces decision complexity by separating strategic targeting from tactical execution, but it does not eliminate this retraining dependence. We investigate whether frozen, zero-shot large language models (LLMs) can provide retraining-free control in hierarchical cyber defense and how performance changes as LLM control is extended from planning to execution. We formulate a controller-agnostic planner-executor hierarchy in which the planner selects a subnet to defend over a fixed horizon and the executor selects defensive actions within that subnet. Using the high fidelity Cyberwheel environment, with its built-in automated red team agent mapped to the MITRE ATT&CK framework, we compare RL+RL, LLM+RL, and LLM+LLM configurations using six models ranging from 3B to 70B parameters, including two cybersecurity-specialized models, across small, medium, and large networks. Replacing only the planner with an LLM yields limited gains as network size increases. In contrast, extending LLM control to execution produces notable improvements for sufficiently capable models. For instance, a frozen general purpose 70B model holds successful lateral movement to approximately 1% of steps and attacker impact near zero across all three network scales using the same model weights, while the RL baseline is retrained for each scale. Our results show that sufficiently capable frozen LLMs can maintain strong defensive performance across the evaluated network scales without task-specific retraining, while also indicating that strong tactical execution is important to realizing the benefits of LLM-based control.

The corpus of worldwide news is important for natural language processing, knowledge graphs, large language models, and other technical efforts. Additionally, this corpus is important for understanding the people, places, organizations, and events that interact in real-time every day. This paper compares two news datasets used for these tasks today, namely the Global Database of Events, Language, and Tone (GDELT) and Common Crawl News. Our research highlights the strengths and limitations of each dataset, analyzing their content and coverage. Notably, while GDELT relies on broadcasts, prints, and web news from across the globe, Common Crawl focuses on news sites from around the world gathered through web crawling. Our analysis revealed considerable differences in where the two datasets gather their news sources.

Activation memory, not compute, limits CNN inference on constrained hardware such as microcontrollers. Direct in-place convolution removes the dual-buffer cost, but the memory-optimal formulation of Gural and Murmann assumes valid padding, unit stride, unit dilation, and odd square kernels, and needs a non-sequential traversal costing $2\times$ inference time in transposes. We identify two regimes in which their published closed form does not hold: (1) an under-allocation of exactly $(k-1)C_{in} \bmod (C_{out}-C_{in})$ scalars, active on every convolutional layer of their own deployed network and manifesting as a silent corruption of still-live input; (2) an unbounded overestimate, up to $2{,}432\times$, once the critical leg leaves the output grid. We correct both and generalize to arbitrary stride, dilation, padding, and rectangular kernels. We then propose Right In-Place (RiP) convolution, a bit-identical operation in which every layer reads its input right-aligned in a shared workspace and writes its output left-aligned from index zero. The debt is piecewise affine in the output pixel index, so evaluating its breakpoints in $O(1)$ yields the minimum safe gap without enumerating the output grid, with row-major access preserved. Across $10{,}000$ random layers RiP produced no corruption, and across 84 convolutional layers from 25 architectures it matches the herringbone workspace exactly on 58 and within 5% on 81, using 24.8% less memory than dual buffering on average. Written into TinyEngine's kernels and deployed to a Raspberry Pi Pico 1 and Pico 2, it cuts peak activation memory across eleven MCUNet models by 12.5 to 33.3% at unchanged cycle counts and bit-identical outputs, raising the number of models that fit the Pico 1's 256 KB SRAM from six to nine.

People are increasingly delegating tasks to AI agents, and those agents are increasingly encountering other people's agents over shared resources such as a codebase, a calendar, or a budget. When each agent acts for a different user with different goals, coordination often fails, and the group ends up worse off than if a single agent had acted for everyone. We study this multi-user, multi-agent setting across five frontier models and 77 scenarios in four environments: an API key environment in which agents share a compute budget, a clinic in which they share a calendar, a personal assistant environment in which they share a group order or booking, and a merge queue in which they share a release cutoff. In each scenario, we compare a single agent that serves every user (a coordinator) to a team in which each agent serves one user, with and without a communication channel between the agents. Teams deliver worse group outcomes than the coordinator in every environment: without a channel, they completely collapse in two environments, and even with one, coordination overhead creates substantial gaps. For example, in the personal assistant environment, the coordinator fulfills a targeted user request about twice as often as teams. We identify distinct behaviors associated with this poor group-level performance, including stalling as teams grow, overriding each other's actions, and fabricating claims. We find effective but environment-specific mitigations, such as a team lead, explicit procedural instructions, and a platform check that makes an agent read its peers' messages before committing. We will release the API key, clinic, and personal assistant environments as MAMUBench, comprising 74 scenarios for evaluating multi-user, multi-agent coordination.

Aesthetic image cropping aims to identify the optimal crop of an image in terms of aesthetics and composition. While supervision based on annotated data is fundamental, the field has been hindered by a long-standing problem: existing datasets suffer from (1) human subjectivity and (2) rigid discreteness confined to fixed sampling grids. These flawed annotations not only limit the accuracy and generalization of trained models but also severely distort fair evaluation. To overcome this, we propose to model human cropping preference as a multi-peaked, continuous, and sharp field over the crop space. We introduce the Continuous Preference Field (CPF), which recovers a dense preference landscape from discrete annotations through (1) peak clustering, (2) off-lattice refinement, (3) negative shaping, and (4) field assembly. Based on this, we train CPIC, a VLM-based cropping model optimized via GRPO with the CPF reward, which overcomes template collapse, achieving state-of-the-art performance and exceptional out-of-domain generalization. Finally, to resolve the long-standing benchmark evaluation crisis, we introduce CPICD, a comprehensive recalibration of existing ground-truth boxes. By leveraging the CPF to correct grid-bound artifacts across mainstream benchmarks, CPICD establishes a rigorous and reliable foundation for future cropping research. Extensive experiments and user studies demonstrate the superiority of our CPF, CPIC, and CPICD. Code, model, and data are available at https://github.com/zzqingz/CPIC.

In collaborative foundation model fine-tuning, client data is rarely homogeneous. Instead, clients typically possess unknown mixtures of distinct data distributions, or tasks. Conventional federated learning primarily addresses heterogeneity across clients without explicitly resolving latent task mixtures within each client. We study this setting as compound heterogeneity, where data is heterogeneous both across and within clients. We study adaptation over a common frozen representation and show that, when tasks share the same feature geometry, the optimal model for a client's task mixture under squared loss is a convex combination of the optimal models for its underlying tasks. Thus, a single locally trained model represents the client's overall task mixture, while individual inputs may be drawn from different underlying task distributions. This motivates routing inputs to specialized experts, and we show that, when the task optima form a simplex, task-aligned routing achieves lower risk than any single adapted model for genuinely mixed clients. With access to a small set of task-labeled public samples, we derive a convex program to recover task experts and match them to their corresponding tasks. Our routing analysis shows that effective specialization requires input-dependent expert selection aligned with each client's task mixture. Motivated by this analysis, we propose FedSEE. Across our experiments, FedSEE avoids the negative transfer observed in the evaluated baselines and improves performance by 2.9 points overall and 3.7 points for the worst-served quartile.

In a district-based election, N voters are partitioned into k districts, and each voter votes for one of m candidates. Each district elects a winner using the plurality rule (i.e. the candidate getting the largest number of votes is declared the winner, breaking ties as per some fixed rule), and the overall winner is determined by applying plurality to the district winners; we assume that there is a unique winner amongst the district winners. The margin of victory of such an election is the minimum number of votes that must be altered so that the current winner ceases to be the unique district winner. We study the problem of predicting the winner of a district-based election in the query complexity model, where one has query access to individual votes. The objective is to minimise the number of queries. This setting captures exit polling, where queries correspond to interviewing voters, and is closely related to problems in query complexity and property testing. Assuming that the margin of victory of the election is at least eps N, Dey, Kar and Sanyal (AAMAS 2023) gave algorithms for the case of two candidates with error probability del and query complexity tilde{O}(1/eps^6 log^2 1/del), which improves to tilde{O}(1/eps^4 log^2 1/del) under the additional assumption that district populations are balanced. Our main result is an adaptive randomised algorithm that, for an arbitrary district-based election and any error parameter del, with probability at least 1-del, predicts the winner correctly using tilde{O}(1/eps^2 log m/del log 1/del) queries. In particular, we improve the bounds of Dey et al. for arbitrary district populations and extend their results to any number of candidates. Furthermore, for constantly many candidates, our algorithm nearly matches a lower bound of Omega(1/eps^2 log 1/del) on the query complexity that holds even for two candidates and a single district.

In this paper, we propose Gestalt, a new paradigm of large multimodal model built around multimodal interplay. Despite rapid advances, large multimodal models are reaching a bottleneck: existing approaches focus primarily on accommodating additional modalities while overlooking the distinct characteristics of each modality and the relations among them. Motivated by the multistage property of human multisensory perception, we propose a multimodal interplay pyramid that organizes multimodal modeling as a progression from modality-specific processing, through cross-modal alignment, to deeper multimodal integration. Guided by this pyramid, Gestalt adopts a unified discrete diffusion framework and an interplay-partitioned architecture, with learnable interplay tokens mediating cross-modal exchange and integration. The pyramid also structures its data organization and training strategy. Strong performance across image generation, multimodal understanding, and text-only evaluation shows that Gestalt significantly improves cross-modal integration while preserving modality-specific information, effectively harnessing the strengths of diffusion-based multimodal models and offering a promising path toward unified multimodal intelligence. Project page: https://GeWu-Lab.github.io/Gestalt.

A common approach to world-model simulation for vision-language-action (VLA) systems is to predict future RGB observations and then re-encode them into policy inputs, introducing an indirect interface between simulation and downstream policy execution. We instead investigate whether world dynamics can be modeled in a compact, policy-oriented state derived from VLM visual tokens. A key challenge is that raw VLM visual tokens are high-dimensional, making efficient and accurate autoregressive dynamics modeling challenging. To address this, we introduce Token-World, an action-conditioned world model that compresses VLM features into a compact token state, learns future dynamics in this reduced space, and maps predicted states back to the original policy-facing representation for downstream use. Across manipulation benchmarks, Token-World improves open-loop feature fidelity and policy-action consistency over recent world-model simulators, with slower degradation over long rollout horizons. In closed-loop evaluation, its simulated policy performance correlates more strongly with reference policy performance than Ctrl-World ($r=0.794$ vs.\ $0.583$), while requiring lower simulation latency. Ablations further show that compact-representation design and dimensionality substantially affect future-state prediction. Code will be available at https://chuyaofu.github.io/Token-World/.

Multi-reward reinforcement learning trains large language models to satisfy multiple behavioral objectives simultaneously. Reward-wise normalization, as used in GDPO, preserves reward-specific relative information within rollout groups, but different objectives can still exhibit uneven learning progress. We study this behavior through advantage energy, the sum of a reward's squared advantages over a batch. Under idealized GDPO normalization, we show that this energy is proportional to active-group density: the fraction of rollout groups in which the reward provides nonzero relative advantages. This reveals a residual batch-level signal imbalance and provides a basis for calibrating reward contributions. Based on this relation, we propose Density-Aware Reward Aggregation (DARA). We derive an inverse-square-root density correction that gives greater weight to signals from less frequently active rewards. DARA computes its weights from each rollout batch, adapting to changes in reward activity throughout training without modifying the underlying policy optimization objective. Experiments on tool calling and mathematical reasoning show that DARA learns the targeted behaviors faster than GDPO, reaching high format compliance in up to 26% fewer training steps on tool calling and near-saturated length compliance in up to 65% fewer steps on mathematical reasoning, while remaining competitive in final performance. Our code is available at https://github.com/zhaihaotian/DARA.

Query-aware keyframe selection enables multimodal large language models (MLLMs) to process long videos using only a small set of question-relevant frames. Existing score-based methods, however, typically search within a fixed, uniformly sampled candidate pool, preventing evidence outside this pool from ever being selected. Given a limited relevance-scoring budget, the key challenge is to allocate evaluations adaptively to promising frames while continuing to explore underrepresented temporal regions. We introduce FORTE, a training-free framework that addresses this challenge through two stages: adaptive relevance scoring and global keyframe optimization. Starting from sparse, uniformly distributed observations, our efficient Gaussian-process relevance predictor estimates relevance for unscored frames, exploiting temporal locality and the approximately banded kernel structure to reduce the core computation from cubic to linear time in the number of frames for fixed bandwidth. The scoring stage then selects which frames to score next by balancing predicted relevance with temporal coverage, prioritizing promising regions while also exploring less-represented parts of the video. The optimization stage selects the final keyframes by maximizing an objective that jointly captures measured relevance and temporal coverage. We derive an exact algorithm that leverages the logarithmic coverage structure to identify the optimal subset of the scored candidate pool in time linear in the pool size, for a fixed final-frame budget. Experiments on four long-video question-answering benchmarks show that FORTE achieves the highest observed mean accuracy among the compared selectors under every tested scoring budget. Further evaluations demonstrate its consistent effectiveness across different relevance scorers and downstream MLLMs.

Large reasoning models (LRMs) have achieved substantial improvements in solving complex mathematical problems, but often produce lengthy, repetitive, or erroneous reasoning trajectories. In this work, we introduce a new interpretability framework, SLIDER, to evaluate the quality of the reasoning process. SLIDER leverages an emerging body of work from information theory called Partial Information Decomposition to disentangle the information about the final answer between two consecutive reasoning steps into non-negative components: unique information (in preceding steps or current step), redundant information, and synergistic information. Building on this decomposition, we propose the *Step-wise Repetitive Reasoning Index (Step-RRI)*, a theoretically grounded measure that assesses whether the answer-relevant information in the current step $S_i$ is predominantly redundant with the past steps $S_{<i}$, relative to its unique and synergistic contributions. To evaluate the effectiveness of Step-RRI in detecting repetitiveness, we apply SLIDER to the redundancy class of the PRMBench dataset where Step-RRI improves step-level redundancy identification accuracy by over $10$ points compared to embedding-similarity and information-gain baselines. Next, we define *Trajectory-RRI*, an aggregate measure of repetitiveness for an individual reasoning trajectory. To demonstrate its practical relevance, we show that average Trajectory-RRI strongly correlates with actual reasoning length across QwQ-32B, DeepSeek-R1-Distill-Qwen-32B, and GPT-4.1, motivating its use as a signal for improving reasoning efficiency. Finally, we introduce *Trajectory-RRI-guided data selection for fine-tuning*, demonstrating that selecting training data based on Trajectory-RRI can improve a fine-tuned model's reasoning efficiency while largely preserving its task performance.

Full detector simulation and reconstruction of collider events are projected to become major bottlenecks at the High-Luminosity Large Hadron Collider, motivating the development of fast, ML-based surrogates. At the same time, LLMs have driven fast progress in generative discrete modeling: autoregressive transformers trained on tokenized data now represent the state of the art across a range of generative tasks. We extend the discrete modeling paradigm by introducing a particle-level generative model trained on residual-quantized full-event data. We demonstrate the ability of this model family to perform conditional generation from detector-stable particles; we study its scaling behavior across a range of dataset and model sizes, characterize the effects of repeated data exposure and demonstrate that token-level loss systematically predicts downstream physical fidelity. These results provide an empirical framework for scalable collider full-event generation based on residual-quantized representations.

Large language models are characterized by three key properties: capability, alignment, and faithfulness. Prior work studies the tradeoffs between capability and alignment, and between capability and faithfulness, but a third tension remains underexplored: the alignment-faithfulness conflict. We show that aligned models systematically deviate from their inputs on unsafe or sensitive content without disclosing the modification, a failure mode we call alignment-induced unfaithfulness (AIU). Unlike capability-driven unfaithfulness, which comes from errors in knowledge or reasoning, this is induced by post-training mechanisms that override adherence to the input. We introduce FaithConflict, a controlled dataset isolating both conflicts, and two complementary taxonomies: behavioral (B1-B8) and chain-of-thought reasoning (C0-C6). Across models, AIU increases with scale and more sharply than capability-driven unfaithfulness, a reverse scaling law; intermediate checkpoints show it is amplified during post-training, with DPO the stage at which the gap both grows most and becomes least visible. Prompting-based mitigation does not resolve it, revealing a capability-alignment-faithfulness trilemma in the design and evaluation of LLMs.

We study algorithms for detecting induced subgraphs corresponding to fixed pattern graphs in host graphs. We show that at least five of the 21 connected graphs on five vertices can be detected in time roughly the product of the number of vertices and the number of edges, and that at least 65 of the 112 connected graphs on six vertices can be detected in time nearly quadratic in the number of edges. We also give algorithms for detecting induced paths and cycles on seven vertices, running in time roughly the number of vertices times the square of the number of edges. Our main technical tool is a generalized notion of tree decomposition width, called (p, q)-width. It yields algorithms whose running times depend on both the number of vertices and the number of edges, and are never worse than existing bounds. Whenever the host graph has fewer than roughly quadratically many edges in its number of vertices, our bounds are strictly faster. For some patterns, including the seven-vertex cycle, our algorithms are optimal under standard complexity-theoretic assumptions. We further develop this approach using pattern-based polynomials that exploit the structure of tree decompositions, not just their width. This gives algorithms for detecting induced paths and cycles on an even number of vertices in bipartite graphs, running in time roughly the (k-1)-th power of the number of edges for paths on 2k vertices, and that same bound times the number of vertices for cycles on 2k vertices. These are faster than the best known algorithms for general graphs.