Interpretable brain-computer interface classifiers that generalize across subjects without calibration remain an open challenge. We test whether prototype-based cross-attention can provide competitive, interpretable event-related potential (ERP) classification under deployment-compatible conditions. We propose ERP-XTTN, a cross-attention architecture that routes input EEG patches to fixed difference-wave prototypes via query-key-only cross-attention with no value projection, so classification depends entirely on attention routing and attention faithfulness is structural rather than post-hoc. Prototypes are derived automatically from extrema in the training-fold difference wave. We evaluate across three public sources (BNCI Horizon 2020, HRI Cursor, and ERP CORE) spanning eight ERP components (ERN, LRP, ErrP, N170, P300, N2pc, MMN, N400), using leave-one-subject-out (LOSO) evaluation with causal filtering at two channel counts (3-channel and full montage), against EEGNet and xDAWN with Riemannian geometry (xDAWN+RG). The mean gap between the best baseline and ERP-XTTN was .018 AUROC at 3 channels and .034 at full montage, arising from two largely distinct sources: a temporal-flexibility cost relative to EEGNet and a spatial-exploitation cost relative to xDAWN+RG, the latter driven by signal-to-noise ratio at full montage. Beyond accuracy, the transparent routing reveals cross-subject signal structure that black-box models cannot: false positives resembled true positives more than true negatives did, indicating that classification errors are neurophysiologically explicable. ERP-XTTN generalizes across diverse ERPs under causal, calibration-free conditions with a small interpretability cost at minimal montages. To our knowledge, this is the first epoch-level LOSO benchmark on ERP CORE.

In this manuscript, we propose and analyze hierarchical Kolmogorov--Arnold neural network architectures employing radial basis functions as activation functions for approximating deterministic functions and random field models. Specifically, we develop a hierarchical radial-basis-function Kolmogorov--Arnold network (hierarchical RBF-KAN) for multidimensional deterministic function approximation and a hierarchical radial-basis-function stochastic Kolmogorov--Arnold network (hierarchical RBF-SKAN) for random field learning. From a theoretical perspective, we establish universal approximation results for both architectures. In particular, we derive quantitative approximation estimates for the hierarchical RBF-KAN, showing that the proposed framework has the potential to partially alleviate the curse of dimensionality in learning high-dimensional functions by reducing the effective dimensionality of the approximation problem. Furthermore, we show that the hierarchical RBF-SKAN can approximate random field models under the Wasserstein-2 metric. Empirically, we show that our proposed radial-basis-function-based neural network structure could effectively learn multivariate functions and random field models.

Accurate registration of CAD models to CT scans is essential for establishing ground truth geometry in volumetric imaging. Obtaining reliable object masks is of growing importance in machine learning settings; as recent architectures grow more capable, huge datasets are required to fully utilise their capabilities. Traditional intensity-based methods fail when CT grayscale values lack calibration references, while point-based algorithms (e.g., ICP, RANSAC) require feature correspondence unavailable between idealized CAD geometry and noisy volumetric CT data. We propose a two-stage geometric registration method for cylindrical objects (ionization chambers) that takes advantage of the distinctive geometric features of the objects. First, we estimate the 3D rotation axis by detecting elliptical cross-sections across CT slices, fitting ellipses to edge-detected contours, and performing PCA on the fitted ellipse centers after RANSAC outlier removal. Second, we voxelize the CAD model, orient it along the detected axis, and maximize volumetric overlap with the CT scan through translational adjustment. This approach achieves robust registration with tilt and orientation errors below $0.1^\circ$ without intensity calibration or feature matching. Once registered, the aligned CAD model provides ground truth geometry for applications including machine learning-based object localization and automated analysis in industrial CT workflows.

Older adults are particularly susceptible to falls following perturbations during standing, such as forward loss of balance. Back support devices that assist trunk extension may help mitigate fall risk by preventing excessive trunk flexion. Previous studies have investigated heavy back support devices; however, these systems often introduced adverse effects on stability due to their added mass, which shifted the body's natural center of mass unfavorably. In contrast, lightweight passive devices have shown limited benefits, as they can generate only modest assistive forces during the relatively small trunk flexion associated with forward balance loss. In this study, we evaluated the effects of a lightweight semi-active soft back support device on postural stability following standing perturbations. Our device combines an active element (a pneumatic artificial muscle) in parallel with a passive elastic band. The active element rapidly provides assistive force following a perturbation, overcoming the limitations of passive devices. Experiments conducted with five healthy individuals demonstrated that the semi-active device significantly reduced whole-body angular momentum and increased the margin of stability, indicating improved balance recovery performance. These results highlight the promise of semi-active soft wearable robots as an effective and lightweight strategy for fall prevention during standing perturbations.

Normalization layers such as BatchNorm and LayerNorm have long been considered essential for stable training in deep networks. This work demonstrates that they can be fully replaced by a single learnable activation mechanism. We identify a plasticity suppression effect induced by standard normalization: learnable activation parameters rapidly lose adaptability when paired with normalization layers. Motivated by this observation, we introduce SALU (Saturated Adaptive Linear Unit), \[ \operatorname{SALU}(x;a,b) = \frac{a x}{\sqrt{1 + a b x^2}},\quad a>0,\; b>0 \] a bounded, learnable activation that provides intrinsic signal stabilization without relying on batch statistics or external affine parameters. Building on SALU, we propose SaluNet, a paradigm grounded in total plasticity: SALU replaces normalization layers, while SWALU and GALU replace standard activations. With ResNet-18, SaluNet-C-18 achieves 97.35\% on CIFAR-10 and 83.25\% on CIFAR-100 without normalization, maintaining 93.44\% and 76.23\% at batch size 1 where normalized architectures fail. For transformers, SaluNet-T improves over LayerNorm-GELU from 90.92\% to 91.01\% on CIFAR-10 and from 66.54\% to 68.10\% on CIFAR-100. SaluNet-C-50 reaches 78.67\% Top-1 on ImageNet-1K at $224\times224$, and $79.23\%$ at $288\times288$. These results suggest normalization layers suppress total plasticity, a property biological neurons inherently possess, enabling deep networks to learn effectively.

Language-model unlearning updates a trained model to behave as if it had not seen selected training examples, while preserving utility and avoiding costly retraining. Existing approaches typically fine-tune the pretrained model with a fixed training budget and select the final model afterwards by evaluating several saved checkpoints on downstream validation data. Two sources of unnecessary computation limit scalability: training beyond the desired forget-retain trade-off, and checkpoint selection that requires extra storage and repeated evaluations. To address these limitations, we introduce MArgin Self-Correction (MASC), an efficient unlearning method with an online stopping rule that does not require downstream evaluation. Given a text sequence to be forgotten, MASC actively reduces the logit gap between the original next token and the most likely alternatives. It outputs a final model once this gap is small on average over a sufficiently large proportion of token positions across all forget sequences. On TOFU, MUSE News, and MUSE Books, MASC achieves a competitive forget-retain trade-off at a fraction of the computational cost of existing baselines. We further observe that as we increase model size (a.k.a. number of parameters), the trade-offs improve for both MASC and SimNPO -- the forget metrics remain comparable while retain utility increases.

Visual posters are a compact medium for communicating dense information, yet progress on automatic poster generation remains difficult to measure because existing evaluations are often restricted to paper-only inputs, narrow domains, or surface-level visual similarity. We introduce Any2Poster Bench, a benchmark for any-source poster generation that evaluates systems across eight input modalities--PDFs, URLs, PPTX, DOCX, Markdown, LaTeX, notebooks, and videos--and five content domains. Any2Poster Bench pairs each source with quiz-based probes of verbatim factual retention and interpretive understanding, together with VLM-based judgments of visual quality, layout, readability, content completeness, and logical flow, enabling reproducible assessment of both information fidelity and visual communication. To instantiate and validate this benchmark, we further present Any2Poster Agent, an end-to-end reference agent that parses heterogeneous sources, organizes salient content, plans poster layouts, renders posters, and iteratively refines them using visual feedback. On Any2Poster Bench, Any2Poster Agent achieves 87.25% average accuracy across input modalities and 87.28% across content domains. On PaperQuiz-style evaluation, where prior paper-to-poster agents are directly comparable, Any2Poster Agent improves over PosterAgent-4o from 51.06-51.33% to 72.58% overall accuracy and from 116-121 to 145.16 in density-augmented score. Together, Any2Poster Bench and Any2Poster Agent provide a reusable evaluation resource and a competitive baseline for studying multimodal, domain-general poster generation.

Current research primarily focuses on model performance, while comparatively less attention has been devoted to uncertainty estimation, particularly in settings where LLMs are increasingly used to generate annotated data. We introduce a framework combining conformal prediction with Collaborative Filtering-style annotators' representation to model LLM behavior in relation to human annotators and to analyze patterns of agreement and disagreement. Using Non-Conformity Scores, we introduce the Ghost Prediction metric and the Ghost Annotator representation to quantify cases in which model predictions diverge from all available human annotations. We compute cosine similarity measures to explore differences in model behavior across sociodemographic axes. We evaluated four LLMs of different size and families across four content moderation datasets. Our finding shows that while we find that all models uncertainty increases with annotator disagreement, larger models tend to be more confident in the classification of texts that are not aligned with any human annotation. Finally, the Ghost Annotator framework reveals a consistent and robust pattern of demographic misalignment, suggesting a structural bias likely rooted in pretraining corpora.

Multi-turn user-facing agents must infer user intent from incomplete requests, collect missing information through dialogue and tools, and execute valid actions. A training trajectory records this process as an interleaved sequence of user messages, agent responses, tool calls, etc. Synthesizing sufficiently complex trajectory has become a central route to train agents: existing pipelines often increase difficulty by composing multiple user requests into longer tasks, producing write-intensive trajectories that train sequential execution. We argue that a single write decision can itself be difficult when the agent must gather and compare substantial read-tool evidence before its arguments become identifiable, a challenge that write-intensive data alone cannot address. Guided by this insight, we propose WRIT (\uline{W}rite-\uline{R}ead \uline{I}ntensive \uline{T}rajectory Synthesis), a pipeline for synthesizing multi-turn agent training trajectories along two complexity axes: the number of write decisions in a task and the evidence burden of each individual decision. WRIT first generates write-intensive and read-heavy tasks. It then diversifies user behavior instructions to reflect realistic conversational variation, and finally simulates agent-user interactions in an executable environment to produce complete training trajectories. The resulting data trains agents not only for longer task execution, but also for robust, evidence-grounded decision making under high information load. With only 2K synthesized trajectories, a 4B model trained on WRIT outperforms GPT-5.1 no-think on $τ^2$-bench and substantially reduces inference-time token usage, showing that compact SFT data can convert part of expensive test-time reasoning into efficient agent behavior.

Breast cancer recurrence, a leading cause of long-term mortality among survivors, requires timely and accurate risk assessment to guide follow-up care and treatment planning. Traditional predictive models, often limited to either structured or unstructured data alone, struggle to capture the full clinical context. This study examines the impact of integrating multi-modal clinical data, including treatment records, pathology reports, and clinician notes, on recurrence prediction. By integrating a rule-based regular expression extraction mechanism with a rigorous precedence-based conflict reconciliation strategy, our approach effectively recovers definitive tumor characteristics from free-text pathology narratives to augment structured records. We also benchmark performance against commonly used feature sets from prior breast cancer studies to assess the added value of multi-modal integration. Single-source and multi-modal inputs are evaluated across a range of machine learning models. Results show that multi-modal integration consistently improves predictive accuracy compared to single-modal methods.

The effectiveness of a soft wearable back-support device in enhancing postural stability was investigated under trip-like perturbations using two experimental paradigms: perturbed standing and perturbed walking. Healthy subjects completed trials under three different back-support conditions: no device, device worn with low stiffness, and device activated with high stiffness. Whole-body stability was quantified using the minimum Margin of Stability (MOS) at the point of maximal instability. Results demonstrated increased MOS during device use, indicating enhanced postural stability. In standing, MOS increased significantly with device stiffness, whereas in walking, both device conditions improved MOS relative to no device but did not differ significantly from each other. These findings highlight the potential of soft wearable back-support devices with adjustable stiffness to improve reactive balance control against external perturbations, with important implications for fall prevention. Future research should explore personalized stiffness optimization and evaluate efficacy in populations at elevated risk of falls.

Reward guidance algorithms steer a learned generative process toward the reward-tilted measure at inference time. While empirically powerful, these methods are prone to reward hacking: the guided model over-optimizes the reward at the cost of fidelity to the learned distribution. Prior work has attributed this to the complexity of neural reward functions or implicit biases in diffusion training, but its fundamental origins remain poorly understood. We show that reward hacking arises from an approximation made in most practical implementations of reward-guided diffusion -- finite-particle plug-in estimation of the Doob h-function -- even in the simplest non-trivial settings of Gaussian and Gaussian mixture targets with quadratic rewards. In closed form, we isolate two distinct failure modes of the plug-in estimator: it leads to reward hacking within each mode and it cannot select high-reward modes. We propose a closed-form reward damping schedule that corrects the within-mode bias with no additional compute, and clarify the role of best-of-n sampling in compensating for the mode selection failure. Experiments on Gaussian mixture targets, a 2D checkerboard, and FLUX.1 text-to-image generation confirm that our theoretical insights carry over to practical settings.

Informed sampling techniques accelerate sampling-based motion planners by focusing the search on promising regions of the state space, yet most existing methods rely on Euclidean heuristics that become inadmissible under configuration-dependent Riemannian metrics. While scalar eigenvalue bounds restore admissibility by uniformly scaling the Euclidean distance, they discard the directional structure of the metric, producing overly conservative informed sets. We propose a matrix-valued admissible heuristic that exploits the Loewner order on symmetric positive definite matrices to compute the tightest constant lower bound on the metric tensor while preserving its full directional structure. The Cholesky factorization of this bound defines a linear map to an isotropic Euclidean space in which the Riemannian informed set reduces to a standard prolate hyperspheroid, enabling direct, rejection-free sampling using existing algorithms. Experiments on manipulation tasks with a 6-DoF UR5, 7-DoF Franka, and 14-DoF PR2 under three distinct Riemannian metrics show that our heuristic produces consistently tighter informed sets than both the Euclidean and scalar eigenvalue bounds, accelerating convergence across multiple state-of-the-art asymptotically optimal planners.

Randomized smoothing (RS) uses a smoothed classifier to provide architecture-agnostic certificates of $\ell_2$ classification robustness, but its dependence on per-input Monte Carlo (MC) sampling undermines its use in real-time systems. We argue that this cost is structural rather than fundamental, such that it can be significantly reduced by sharing information across the deployment stream. We introduce RRISE, an RS framework that compresses certification into a single forward pass through a learned surrogate. RRISE trains the surrogate against precomputed MC class-count targets via a soft-label cross-entropy loss and converts surrogate predictions into provably conservative certified radii through a one-time conformal calibration step. The resulting certificate is deployment-verifiable: whenever the calibrated radius is positive, the surrogate's prediction provably matches the smoothed classifier's and the smoothed classifier is constant on a ball of that radius around the input. Across image classification benchmarks, RRISE matches fixed-budget MC certified accuracy within $0.84$ percentage points while replacing up to $10^4$ noisy base-model evaluations per query with a single surrogate forward pass, recouping MC training cost after $\approx 10^5$ deployment queries. On CIFAR-100 and Tiny ImageNet, where the only prior offline-surrogate method collapses, RRISE achieves $1.23$ to $1.91\times$ higher certified accuracy, establishing efficient randomized smoothing as a practical path to certified robustness in repeated-deployment settings.

Coding-agent benchmarks evaluate whether a single uninterrupted agent can resolve a repository issue. Real software work is messier: tasks are interrupted, reassigned, reviewed, and resumed from partial states left by another agent or engineer. We study this missing dimension through \emph{handoff debt}: the rediscovery cost imposed when a predecessor's work is opaque or incomplete. Our takeover protocol interrupts a coding agent at deterministic handoff points, freezes the repository, and evaluates successor agents under four handoff views: repository state only, raw trace, summary notes, and structured notes. Across 75 source tasks, the protocol generates 181 handoff-point tasks and 724 takeover runs per successor model. Across three successor models, context-bearing handoffs reduce median agent events by 20--59\% and cumulative prompt tokens by 42--63\% relative to repository-only takeover. Solved-rate effects are smaller and model-dependent, but efficiency gains are consistent. These findings suggest that coding-agent evaluation should report not only whether a task is solved, but also how costly that work is for another agent to resume.

Large reasoning models improve performance by generating extended chain-of-thought (CoT) reasoning, but this behavior becomes inefficient when applied to LLM agents. Current LLM agents often generate verbose textual reasoning at every decision step and allocate reasoning effort nearly uniformly across turns, leading to substantial inefficiency in multi-turn agentic trajectories. We propose Adaptive Latent Agentic Reasoning (ALAR), a dual-mode framework that uses compact latent reasoning for routine turns and selectively escalates to explicit chain-of-thought when deeper deliberation is needed. ALAR learns latent reasoning by using the agent's actions as supervision anchors and is further optimized to use latent reasoning when it is sufficient for task success and reserve explicit CoT for harder decisions. Experiments on agentic search and tool-use benchmarks show that ALAR maintains comparable or better task accuracy while substantially reducing generated tokens by up to 43.6% in search and 84.6% in tool use. These results demonstrate that ALAR improves the accuracy-efficiency trade-off of LLM agents by reducing unnecessary textual reasoning while preserving explicit deliberation for harder decision steps.

When does multi-agent debate help data cleaning, and when does it hurt? Across three benchmarks, four model families, and over 6,000 task-condition pairs, we find debate's effect reverses sign: it degrades generation across all four models (-1.6 to -15.5pp) through critique-induced confusion (CIC), hallucinated Critic feedback that the Generator accepts uncritically, yet improves error detection (+27.4pp F1, d=1.0). We derive a debate benefit condition: debate helps when the probability of rescuing a wrong output (Critic verification odds weighted by fixability) exceeds the probability of destroying a correct one. A factorial experiment proves adversarial separation is essential: self-verification with identical tools fails, while a separate Critic with code-execution grounding and evidence-gated generation produces the first debate configuration to significantly exceed single-agent on a generative task (+5.3pp, p<0.05). The condition correctly predicts all nine task types and generalizes with zero false positives across 19 published comparisons in seven domains.

AI-Driven Research Systems (ADRS) -- systems coupling LLMs with automated evaluation to discover algorithms, proofs, and designs -- are being optimized and adopted across domains, but the tools to analyze them have not kept pace. ADRS performance depends on component interactions that are poorly understood, expensive to explore, and (as we show) not well captured by standard convergence guarantees. These guarantees rely on structural assumptions that do not hold under the ADRS process we formalize. We introduce GAMBLe, a framework that decomposes ADRS behavior into four parameters (generator $G$, assessor $\mathcal{A}$, discovery mechanism $\mathcal{M}$, budget $B$) and one compositional object, the effective landscape $L_{\text{eff}} = \mathcal{A} \circ G$, which reveals that distinct generator-assessor pairs induce structurally different per-problem optimization landscapes. We exercise the framework on 760+ replicated runs (>46,000 iterations) spanning generators from single LLMs to dynamically-adaptive ensembles, mechanisms from greedy selection to co-evolutionary meta-search, and three NP-hard problems whose assessors range from continuous scoring to cliff functions. The experiments reveal no total ordering of generators or mechanisms: frontier models can underperform open-source alternatives and the simplest mechanism sometimes outperforms state-of-the-art meta-search. Results show that even under limited budgets (60 iterations per run), the right component choices can improve performance by 13-67% and search efficiency by 6-39x.

The rise of Large Language Models (LLMs) has enabled agentic AI capable of complex reasoning and tool use; however, deploying such autonomy in pervasive computing environments remains challenging due to the strict memory and energy constraints of embedded microcontrollers. Existing frameworks typically assume server-class resources or continuous connectivity, leaving a gap for deeply embedded systems. This paper proposes a modular reference architecture for Embedded Agent Systems that bridges the divide between deterministic real-time control and agentic intelligence. We introduce a tiered design that decouples On-Device Agents - executing highly compressed neural networks and rule-based logic for low-latency, privacy-critical tasks - from Cloud-Augmented Agents that leverage Small Language Models (SLMs) for higher-level reasoning and planning. A key contribution is the integration of a cross-cutting Governance Layer, ensuring observability, policy enforcement, and safety across distributed fleets of autonomous devices. Rather than presenting purely empirical benchmarks, we analyze architectural design principles and trade-offs regarding latency, energy, and reliable execution in resource-constrained environments.

Catastrophic forgetting is often framed as a representational problem: after sequential training, a model appears to lose the features that supported performance on earlier tasks. We challenge the stronger form of this view. Across controlled continual-learning settings, we find that a significant portion of apparent forgetting can be attributed to interface drift between internal stages rather than permanent erasure of task-relevant computation. We study this phenomenon through a stitched evaluation protocol that combines early computation from a post-update network with late computation from its predecessor, optionally mediated by a compact, task-specific transport key. We describe transport keys at a systems level as compact interface-alignment operators estimated from a small set of paired anchor activations and evaluated through model stitching. On split CIFAR-100 with a ResNet-style network, transport keys recover most of the original Task A performance after sequential training on Task B. On a compact vision transformer, we observe a similar recovery pattern. These results suggest that continual learning may require better mechanisms for indexing and re-accessing latent computations, not only methods that prevent weight change.

How can a population of agents self-orchestrate and self-adapt into stronger collective intelligence without centralized control? Inspired by Friedrich Hayek's economic theory of decentralized coordination in markets, we study this question through an agent economy in which agents compete via auctions for the right to act, exchange payments, and accumulate wealth from environmental rewards. These simple economic signals induce decentralized credit assignment, driving planning without global orchestration or explicit communication protocols. The population evolves through economic selection: effective agents accumulate wealth and are mutated via exploitation, while ineffective ones go bankrupt and are replaced via exploration. We show that, initialized with weak agents, the economy produces emergent multi-step reasoning strategies and outperforms stronger monolithic baselines across five agentic tasks, including mathematical reasoning, financial research, scientific research, accelerator design, and distributed-system optimization. We further provide theoretical insights into how economic dynamics shape agent behaviors, linking local incentives to long-term global performance. Our results suggest a new path to multi-agent intelligence: rather than engineering coordination, we can design decentralized incentive structures under which it automatically emerges.

Zeroth-order (ZO) optimization is a memory-efficient alternative to backpropagation for fine-tuning large language models, but its deployment is limited by the high variance of gradient estimation. We propose GRZO, a Group-Relative Zeroth-Order optimizer that draws one pseudo-independent perturbation per mini-batch example and aggregates the per-example losses through group-relative normalization, raising the effective gradient-direction count from one to the batch size at no additional forward cost while preserving inference-level memory. We prove that GRZO is directionally unbiased with variance shrinking proportionally to the batch size, yielding a tighter nonconvex convergence bound than MeZO. Across RoBERTa-large, Llama3-8B, and OPT-13B over multiple tasks, GRZO improves average accuracy on Llama3-8B by $+3.0$ over MeZO at $23\%$ lower peak GPU memory; as a drop-in replacement for the MeZO core, it lifts sparse, low-rank, and quantized ZO variants by $+6.0$ on average.

Deep neural networks learn representations where individual features often lack interpretable meaning; a single neuron may activate for scattered, unrelated inputs. We introduce coherence, a geometric property inspired by neural coding in the brain, where neurons like grid cells and head direction cells respond to contiguous regions of state space. A non-negative matrix is coherent if each row (sample) attends to geometrically clustered columns (features) and vice versa, and in addition every sample is well described by some feature and every feature is needed by some sample. We prove that coherent matrices induce a bounded interleaving between the Vietoris-Rips filtrations of samples and features, guaranteeing that both spaces share compatible topological structure. This geometric constraint facilitates interpretability. For example, if data lies on a circle, coherent features must tile that circle into contiguous arcs. We introduce Coh, a differentiable objective function based on Fréchet variance that enforces coherence during training. Unlike sparsity, which bounds how many samples a feature activates on, coherence bounds which samples, requiring geometric connectivity rather than only rarity. This yields not just interpretable features but an interpretable feature space. We validate Coh in an auto-encoder using synthetic and rotated MNIST datasets and in a token embedding of BERT using language data.

Accurate translation from Natural Language to First-Order Logic (NL-to-FOL) underpins neurosymbolic AI systems and Natural Language Inference (NLI), making the quality of NL-to-FOL benchmarks essential -- yet these datasets have never been rigorously audited. Our first contribution is to present a systematic human inspection of the validation split of \textsf{FOLIO} and a subset of \textsf{MALLS} test instances, finding that approximately 39% and 36% of entries, respectively, contain incorrect FOL formalizations (i.e., ground truth labels), with additional rates of ambiguous NL sentences (16.4% and 48%) and incorrect NLI labels in \textsf{FOLIO} (8.4%). Our second contribution is to develop and release corrected ground truths for such datasets, showing that annotation errors distort model evaluation on a reference benchmark task: testing three state-of-the-art LLMs (Gemma~4 31B-it, Qwen3-30B-A3B, and GPT-4o-mini) with the corrected ground truths yields accuracy gains from +9 to +22 percentage points. Motivated by these findings, we propose an LLM-based framework to support humans in manual reviewing NL-to-FOL datasets. By directing reviewers toward the most error-prone instances, we empirically show that it is possible to achieve 90% dataset accuracy after reviewing fewer than 24% of instances, compared to over 70% required by unguided review. We release all human-verified annotations and the code for our framework.

Large Reasoning Models (LRMs) improve performance by generating explicit intermediate reasoning traces through increased test-time compute, yet the assumption that longer reasoning is consistently beneficial remains under-examined. While recent evidence shows that additional reasoning can lead models to overthink, we ask: "Once a model has reached the correct answer, does further reasoning refine the solution, or deviate from it?" To study the dynamics after correctness, we introduce a prefix-level trajectory evaluation protocol grounded in reasoning sufficiency, defining the minimum reasoning budget required for a model to first generate the correct answer. This allows us to disentangle verbose overthinking, where additional reasoning is redundant but harmless, from harmful overthinking, where continued reasoning destabilizes an already-correct trajectory. Starting from multimodal benchmarks, we find that many instances considered reasoning-intensive require surprisingly little reasoning. Moreover, stopping at the first correct prefix improves accuracy over standard reasoning up to 21%, revealing that current models are limited not only by their ability to reason, but also by their inability to stop at the right time. Furthermore, while common efficiency strategies like early stopping substantially reduce verbose overthinking (up to 50%), they fail to mitigate harmful overthinking. Failure analysis reveals that correctness deviations are mainly driven by logical drift and visual reinterpretation. Finally, we show that our findings generalize to language-only reasoning benchmarks, highlighting harmful overthinking as a broader reliability risk. Code available at https://simonecaldarella.github.io/thinking-past-the-answer.

Despite a great deal of prior research into Procedural Content Generation (PCG), relatively little prior work has explored generating enemies for video games. In particular, there is almost no work on generating enemy morphologies, the basic body plan or collision information for in-game enemies, despite the existence of related morphology generation work in robotics. In this paper, we explore three different novel approaches to generate enemy morphologies based on player collision information. We found that each approach provides different strengths and weaknesses, but all had equivalent or better performance than an evolutionary baseline adapted from prior robotics morphology work.

Real-world datasets often contain spurious correlations that are not causally related to the target label. When such correlations dominate the majority of training samples, models tend to rely on them, leading to misclassification of minority samples that do not exhibit the same spurious patterns. While a potential approach is to select subsets of data to better represent the minority samples, this may require access to group labels, which are typically unknown. Furthermore, as we demonstrate, widely used sample scoring functions in the invariant subset or coreset selection literature largely depend on spurious features and therefore fail to accurately capture the importance or difficulty of core, causally relevant features. Accordingly, we propose to mitigate spurious correlations by developing a two-stage sample scoring function that disentangles the learning dynamics of core and spurious features and evaluates their difficulty separately. Based on our proposed metric, we introduce a new algorithm to find and prioritize informative samples both with and without spurious correlations. Extensive experiments demonstrate that a standard ERM model trained on our selected samples achieves superior performance compared to state-of-the-art debiasing techniques, while requiring as little as 10\% of the original training data.

Two-bit weight quantization is attractive for memory-efficient LLM inference, but the standard W2 level set {-2,-1,0,+1} often collapses under aggressive W2A4/KV4 settings. We study the scalar level-set geometry of two-bit weights in a Hadamard-rotated quantization pipeline. Conventional asymmetric W2 substantially improves over the standard level set, indicating that W2A4 failure is not only a bit-width problem but also a reconstruction-level problem. Across all 224 linear modules in each of LLaMA-2-7B and LLaMA-3.1-8B, pretrained weights are already nearly zero-centered, while Hadamard rotation primarily Gaussianizes their standardized shape: excess kurtosis and Q-Q error drop by orders of magnitude. Based on this approximate zero-centered Gaussian-like source model, we propose Qift, a fixed no-zero W2 level set for rotated W2A4/KV4 inference. The main level set is {+/-0.5, +/-1.5}, equivalently {+/-1, +/-3} under a half-scale reparameterization; a power-of-two variant uses {+/-1, +/-4} for sign-and-shift decoded weight application. Qift redesigns the fixed two-bit code-to-level mapping and is training-free, learned-codebook-free, group-grid-free, and zero-point-free, retaining the standard per-channel scale. A scale-invariant ratio analysis identifies an effective inner/outer centroid ratio range of 0.25 to 0.33, explaining why mirror no-zero (MNZ), Lloyd, NF2, and PoT-MNZ perform well while {+/-1, +/-2} does not. On both models, the no-zero level sets consistently improve pure W2A4 perplexity, L-layer mixed W2/W4 perplexity, downstream accuracy, and GPTQ residual behavior over the standard W2 level set. At L=16 mixed precision, they substantially narrow the gap to W3A4 while keeping half of the transformer layers at two-bit precision, giving a simple, source-aware, and deployment-friendly alternative to more complex learned W2 codebooks.

Modeling patient trajectories from longitudinal electronic health records (EHRs) requires reasoning over sparse, noisy, and long-context multimodal sequences. Existing LLM-based multi-agent systems address context length but process patients in isolation, failing to mirror how clinicians leverage accumulated experience from similar prior cases. We present Traj-Evolve, a self-evolving multi-agent system with two complementary evolving mechanisms. First, an Experience Pool (ExPool) acts as a non-parametric memory, indexing rejection-sampled reasoning traces to retrieve similar patients as few-shot contexts. Second, multi-agent reinforcement learning (MARL) via reward-ranked fine-tuning parametrically optimizes inter-agent and agent-memory collaboration. A leave-one-out cross-retrieval strategy unifies the two, aligning training- and inference-time behavior under retrieval augmentation. On a lung cancer prediction task utilizing up to five years of multimodal EHRs, Traj-Evolve outperforms 9 strong baselines on the overall population and a challenging never-smoker population. Analysis of the evolving dynamics highlights three key findings: (1) expanding the ExPool shifts optimal retrieval from diverse to specific samples; (2) under MARL, the manager agent's prediction loss converges quickly while the worker agents' temporal reasoning continues to benefit from more verified patients; and (3) the two mechanisms are complementary on the predicted risk, where ExPool improves specificity while MARL improves sensitivity.

Evaluating vision-language models (VLMs) on medical images requires benchmarks that are clinically grounded, scalable, and controlled for evaluation confounds. Existing public benchmarks are limited in scale, manually annotated, or potentially leaked into VLM pretraining corpora. We present an automated agent-driven pipeline that generates multiple-choice VQA datasets directly from paired private radiology reports and 3D oncology imaging, producing two complementary question types: RADS-style questions deterministically derived from clinician-defined reporting schemas, and radiology report-derived questions generated by an LLM from radiologist findings and verified against the source report. Applied to four in-house cancer cohorts, the pipeline yields an instance-contamination-controlled benchmark without per-question human annotation. Zero-shot evaluation of six VLMs reveals no dominant model and substantial headroom across all cells. A blind ablation reveals that visual reliance is highly dataset-specific: liver Report-derived questions genuinely require the image, while Lung CT is essentially solvable without it - the leading closed model exceeds its sighted accuracy on Lung CT when blinded - indicating that even private clinical data does not guarantee a contamination-controlled read of visual capability. The pipeline is released as an open agent skill for in-house redeployment.