Papers

We present a system that uses a Vision-Language Model (VLM) as a diagnostic agent for adapting a detect-to-track pipeline to a new target domain without access to target-domain labels. Rather than optimizing against annotated metrics, the VLM directly inspects rendered tracking outputs, identifies visual failure modes, and recommends parameter updates through an iterative tuning loop. We first demonstrate that ground-truth-supervised hyperparameter transfer can be brittle. On MOT17->MOT20, applying a source-derived oracle configuration reduces mean HOTA by 0.090, from a target-domain ceiling of 0.357, to 0.267. Without using any target-domain labels, our VLM-based tuner recovers 67.8% of this lost headroom, finishing within 0.029 HOTA of the target ceiling; on the highest-density target sequence, it recovers up to 86.7%. We further show that label-free Bayesian optimization with handcrafted proxy objectives struggles under large domain shifts and can degrade configurations that are already strong. In contrast, the VLM tuner acts selectively: when its visual diagnosis reveals no clear failure mode, it declines to modify the configuration, preserving performance on easy transfers while improving hard ones. Finally, we characterize the conditions under which this approach succeeds, namely, when domain shift manifests through exposed detection-level parameters, versus where it is less effective, such as MOT17->DanceTrack, where the source oracle is already near-optimal.

Employees are increasingly using personally owned generative AI tools such as ChatGPT, Gemini, and Claude for their daily work. This practice is known as Bring Your Own AI (BYOAI), which is a distinct form of Shadow AI in which employee-authenticated personal accounts are used outside of enterprise identity and security controls. Existing frameworks were designed for AI tools managed by organizations, and their coverage does not extend to unmanaged AI tools used with a personal account. In addressing these issues, we developed a governance model through a systematic review of the literature that produces a risk taxonomy and a framework-engagement profile. We also developed a parameterized governance model that measures how much a level of governance maturity reduces residual risk. A five-level maturity ladder is coupled to a technical control architecture through a chain in which the coverage of the control layer influences the security outcomes. Our study of a curated corpus of 30 records (24 research studies and 6 framework documents) indicated that the most prominent categories identified were data exposure and compliance, and framework engagement was inconsistent. Three mutually supporting pillars (technical, governance, and human) were established to support safeguards. Additionally, the results of the model demonstrated that prohibition-based solutions will result in residual risk levels close to those achieved through baseline solutions. Under the specified parameterization, layered control-based solutions substantially reduce modeled exfiltration risk and increase enforceable coverage.

Cryptocurrency markets exhibit extreme volatility and non-stationary dynamics that challenge conventional forecasting methods. Although Large Language Models (LLMs) have shown promise for time series forecasting, the combined effects of adaptation choices remain largely unexplored in financial settings. This study introduces PRICE, a structured approach for adapting LLMs to short-term Bitcoin price forecasting. Built on a 4-bit quantized LLaMA-3 8B model, PRICE investigates how fine-tuning, numerical representation, prompting, inference, and decoding jointly influence forecasting performance. PRICE integrates Parameter-efficient fine-tuning with Low-Rank Adaptation (LoRA), Recursive multi-step inference, Integer-rounded numerical representation, Context-Task-Format (CTF) prompting, and Exact zero-temperature decoding. Ablation studies show that each component contributes to forecasting accuracy and reliability. LoRA enables efficient training on limited hardware, recursive inference improves accuracy, integer-rounded values reduce errors, CTF prompting outperforms Chain-of-Thought, Implicit Chain-of-Thought (iCoT), and few-shot prompting, and zero-temperature decoding improves stability during recursive forecasting. Comparative evaluation against eight transformer-based and time-series foundation models shows that PRICE achieves the lowest forecasting errors on both validation and test sets while maintaining robust performance across evaluation periods. Despite being based on a model primarily pretrained on text rather than time-series data, PRICE achieves competitive or superior performance relative to specialized foundation models. These findings demonstrate that adaptation choices critically determine the accuracy and robustness of LLMs for numerical time-series forecasting.

Continuous sliders are useful only when coefficient changes produce predictable image changes. Yet most diffusion sliders derive their axes from text or learned representations, leaving their scales disconnected from observable image properties. Consequently, we cannot tell in advance which attributes are learnable, compare control strengths directly, or anticipate interference when multiple controls are combined. We propose Measured Sliders, a framework that defines continuous controls through closed-form differentiable image measurements. A common measurement space unifies the pipeline. Before training, an observability test identifies usable supervision. During training, a measurement-guided objective learns target movement while suppressing non-target changes. After training, decoded calibration expresses controls in comparable units of realized image change. Multiple LoRA branches are stored in one checkpoint and composed without training on joint activations. Across SDXL and FLUX.1-dev, the resulting controls are ordered, selective, and composable. On 553 prompts, lighting direction reaches rho = 0.995 and 98.9% monotone sweeps. A five-attribute checkpoint achieves average selectivity 2.59, compared with 1.50 for the strongest baseline, and preserves every requested direction in 96.7% of pair and 86.1% of triple compositions. The observability test also separates every subsequently successful measurement from the failed candidate. Overall, image-space measurement provides a common basis for learning, diagnosing, calibrating, and composing continuous generative controls.

While machine-learning interatomic potentials (MLIPs) have successfully learned potential energy surfaces (PES) and atomic forces, many practical applications, such as vibrational analysis and transition state search, rely heavily on the PES Hessian. Yet standard MLIPs are trained on energy and forces only, and using the Hessian as an additional training objective requires backpropagation whose memory cost grows quadratically with the number of atoms. To address these limitations, we propose two Hessian-derived data augmentation schemes: isotropic Gaussian displacement (UniAug) and normal mode-weighted displacement (ModeAug). Both methods utilize simple Taylor expansions, generating physically grounded molecular structures with highly accurate energy and force labels. By preserving the underlying MLIP architecture and training procedure, this approach contributes to improving energy and force predictions in various benchmarks. Comprehensive evaluations across non-equilibrium and equilibrium datasets demonstrate that our approach enhances model accuracy where reference forces are large and training data are limited.

Autonomous AI agents increasingly select actions in environments whose memory, execution-time, runtime, compute, and operational constraints determine what counts as a suitable plan. We call the absence of this execution context from an agent's planning state substrate blindness. We test this general proposition through numerical code generation, where selected implementation choices and operational consequences are directly observable. Three frontier model configurations–Anthropic Claude Opus 5, OpenAI GPT-5.6-Sol, and Google Gemini 3.7 Flash–generate code for a high-dimensional pairwise Euclidean-distance task either from the task alone or with a 128 MB RAM and 10.0 s wall-time contract. Contract disclosure reduced measured peak process memory in 13 of 14 executable index-aligned task-only versus contract-disclosed comparisons and reduced mean wall time in all three cohorts, making execution up to 3.1x faster. Across the audited corpus, disclosure produced structural code changes including bounded blocking, float32 retention, upper-triangle traversal, and in-place or memory-mapped buffers. At a tighter 96 MB contract, independently sampled contract-disclosed cohorts achieved correct-and-within-budget outcomes of 4/5 for Claude Opus 5, 5/5 for GPT-5.6-Sol, and 3/5 for Gemini 3.7 Flash, compared with task-only outcomes of 0/5, 1/5, and 0/5; cohort mean MaxRSS and wall time were 49-74% and 35-64% lower than their task-only references. These results establish a controlled proof of concept for substrate-aware agent planning: a minimal execution contract induces proactive structural adaptation in generated programs, shifting computation away from unconstrained allocations and substantially improving observed resource-time profiles before execution.

We study latency-optimal file assignment in geo-distributed storage systems modeled as weighted graphs, where edge weights represent communication delays and each node stores one (possibly coded) file. Our goal is to minimize the average time required to retrieve an original file, taken uniformly over all nodes and files. We show that for every fixed number of files $k \geq 3$, computing a latency-minimizing assignment is NP-hard via a reduction from the domatic number problem. On the positive side, we identify natural network topologies that admit uncoded, structured optimal assignments in which, for every node, one can choose its $k$ closest nodes, including itself, so that they store distinct original files. We prove that every weighted tree, certain weighted cycles, and unit-weight graphs with sufficiently large minimum degree admit such assignments. For these graph classes, we provide efficient algorithms to construct latency-optimal file assignments.

Mixture-of-Experts (MoE) architectures provide an efficient paradigm for scaling large language models (LLMs), yet fixed top-k routing activates the same number of expert slots for every token, causing substantial redundant computation. Existing expert-skipping methods often rely on router confidence, calibration data, or additional training, and therefore cannot reliably estimate the actual contribution of routed experts. To this end, we propose ACE, a training-free, calibration-free, and checkpoint-preserving framework for token-adaptive expert skipping in MoE-based LLMs. ACE contains two complementary components: 1) Global Spectral Proxy (GSP), which estimates global transformation capacity from the coupled gate, up, and down projections together with RMSNorm scaling; and 2) Router-Conditioned Refinement (RCR), which constructs expert-specific direction prototypes from centered router weights and evaluates expert responses along routing-preferred directions. During inference, ACE combines both estimates with runtime router gates and skips an expert slot only when both views identify it as low-contribution, while always retaining the top-1 expert. All expert statistics are computed offline, leaving only table lookups and lightweight scalar operations online. Extensive experiments across three MoE-based LLMs and eight benchmarks demonstrate that ACE consistently outperforms existing static and dynamic baselines, with increasingly pronounced advantages under aggressive expert skipping. For instance, at a 50% skipping ratio on Qwen3.6-35B-A3B, ACE reduces WikiText-2 perplexity by 7.96% and improves average downstream accuracy by 4.15 percentage points over the strongest competing method.

Recent reports during the AAAI-27 review cycle highlight the risk of reviewers coordinating bids for reciprocal assignment advantage. Prior work treats bidding, reviewer assignment, and review manipulation as separate stages, leaving the lifecycle effects of collusive bidding unclear. Real-world analysis is further constrained by typically unobservable collusive intent and the lack of counterfactuals for the same conference. Motivated by this gap, we introduce \alg, an end-to-end multi-agent simulacra framework for studying reviewer assignment integrity by holding the conference environment fixed and configuring LLM-driven reviewer agents with honest or collusive policies. We further develop an affinity-guided collusive bidding strategy that uses mutual reviewer-paper affinities to construct collusion rings and select target papers, producing expertise-consistent rather than arbitrarily targeted attacks. Controlled experiments show that collusive bidding more than doubles target-paper capture and that assigned colluders score target papers about two points higher than honest co-reviewers, while conference-wide effects remain comparatively modest. Evaluated bid-phase detectors provide only limited evidence of collusion: in a fixed-triplet detector stress test, native positive-bid graphs are confounded by benign affinity, while a Very-High-only diagnostic view enables precise but low-coverage local recovery.

The struggle of the hour in quantum computing research is achieving effective suppression of the error mechanisms induced by the interaction of external radiation with superconducting quantum devices. Despite the rapid advancements in quantum error correction (QEC) of recent years, radiation-induced faults are yet to be fully addressed. These events are known to be the cause of simultaneous correlated defects in qubits that lie onto a single substrate, ultimately jeopardising QEC code effectiveness. In this paper, we propose to selectively combine substrate-level phonon barriers and QEC interleaving via a planar-mesh tiling algorithm, TETRIS-Q, reaching efficient and effective suppression of radiation events. Our cross-layer solution comes at no extra cost in terms of QEC code execution or decoding time. We model and simulate radiation-induced transient faults over a plethora of barrier and QEC interleaving configurations. Through more than 51 million quantum circuit simulations, we show peak logical error reductions of more than $99.8 \%$, together with an $80\%$ reduction of the observable transient duration with permeable barriers. We find that sparser tiling can reach comparable performance to single qubit tiling, prompting cost reductions of upwards of $87 \%$ in barrier tracing. By leveraging independent QEC code interleaving, we measure up to one order of magnitude average logical error rate reductions without the use of permeable barriers, and up to three orders of magnitude with the joint usage of barriers.

One-shot information theory measures often require an optimization over states, but the form of these optimizers can be complicated or depend on the initial problem in nonlinear ways. In this note, we show that in many instances using the marginal instead of the optimal state is sufficiently good and only changes the result by a small factor. We prove that for the Petz–Rényi and sandwiched–Rényi divergences of order $\alpha\in(0,1)$, replacing the optimizing state on $B$ by the marginal $\rho_B$ results in a multiplicative overhead of at most $1/\alpha$.

Recent progress in multimodal large language models (MLLMs) has fueled significant enthusiasm in their potential to act as autonomous agents for real-world tasks. However, scenarios requiring agents to fulfill users' complex, structured requirements remain largely underexplored. In this work, we examine reasoning tasks under three distinct requirement scenarios: (i) Must-have requirements uniquely determine a unique feasible solution; (ii) Multiple answers satisfy the must-have requirements and are prioritized via the nice-to-have requirements; and (iii) No candidate solution satisfies the must-have requirements, in which case the agent should abstain from generating a response. We evaluate state-of-the-art MLLMs on 3,649 carefully constructed problems that reflect realistic service scenarios, including e-commerce, booking, and map-based or ride-hailing. Our evaluation reveals that existing MLLMs exhibit catastrophic failures in all scenarios. They frequently misinterpret task requirements, violate must-have requirements, and produce invalid solutions. To address this critical gap, we propose First Things First Reinforcement Learning FTF-rl that explicitly optimizes reasoning over multi-priority user requirements. Experimental results show that our method substantially improves the task success rate compared to strong baselines. Moreover, FTF-rl yields general effectiveness on popular logical and mathematical reasoning tasks, including LogicVista, MathVision, and InfoQA. Our findings suggest that enhancing requirement-aware reasoning capability provides a simple yet effective pathway to improve generalization of MLLM agents. Code and dataset are available at https://github.com/claire62/FTF-RL.

Artificial intelligence models are promising for medical diagnosis, but they require large numbers of unbiased data, which in medicine are distributed across hospitals and cannot be centralized to protect patient privacy. Federated Learning (FL) addresses this, since hospitals train one shared diagnostic model while patient data remain local. Training proceeds in communication rounds, in which each hospital trains the shared model locally and returns it to the server for merging by weighted average. This aggregation weight determines whose institutional knowledge shapes the result. Federated averaging (FedAvg) sets it in proportion to local sample count, so a small but informative hospital is permanently assigned a small influence, andl argest clients could dominate the global model even when they are less informative. We propose Federated Dual Reputation Annealing Weighting (FedDRAW), a server-side aggregation method that combines a data-size prior with the cosine similarity between client and global parameters under two coupled annealing schedules. An inner schedule shifts client reputation from the size prior towards similarity. An outer, deferred annealing schedule on the softmax inverse temperature keeps the weighting selective in the early and middle rounds and relaxes it to uniformity at convergence. We evaluate FedDRAW on 12 simulated client-partition scenarios of two chest radiograph datasets (CheXpert and ChestMNIST), against seven federated baselines under identical local training settings. FedDRAW achieved the highest average rank among all eight methods under both AUC and the geometric mean (GM) of sensitivity and specificity, which a Friedman test with Nemenyi post-hoc analysis confirmed to be a statistically significant difference between the methods. Scheduling two signals, rather than fixing the weights by sample count alone, could enable less biased diagnostic models.

Large language models (LLMs) show strong reasoning ability, but their explanations can remain inconsistent, weakly grounded, or difficult to verify. We propose a verifier-guided explainable reasoning framework for transparent educational question answering that combines gold-anchored QLoRA, task-aware symbolic routing, and group-relative RLVR. Qwen2.5-3B-Instruct is first adapted with field-weighted QLoRA supervision anchored to authoritative answers. A lightweight router then assigns logic problems to a FOL/Z3 verifier and physics problems to a formula- and unit aware symbolic solver. Verifier feedback is further used to support candidate evaluation, self-revision, and reward construction during RLVR. Candidate responses are evaluated along three complementary dimensions: P1 for answer correctness, P2 for evidence or unit consistency, and P3 for reasoning depth and explainability. At inference, gold-free self-consistency aggregates multiple candidate responses before an optional question-only physics verifier performs conservative system-level correction. On 438 held-out examples, RLVR increases P3 from 50.68% to 72.20%, while hybrid P1 remains approximately stable at 55.94%. Self-consistency improves model only P1 from 48.86% to 50.23%, with symbolic verification providing the remaining hybrid gain. These results indicate that RLVR primarily strengthens explicit reasoning structure, while symbolic verification complements the neural policy by improving answer reliability at the system level.

The Appropriately Combined Edge-length (ACE) sequence in A-BLiN depends on the zooming dimension $d_z$. This note removes that dependence. The next edge length is selected from the number of cubes that survive the preceding elimination. The resulting Count-Adaptive BLiN algorithm does not use $d_z$ or the zooming constant $C_z$, yet it attains $\widetilde{\mathcal O}_d(T^{(d_z+1)/(d_z+2)})$ regret with $\mathcal O_d(\log\log T)$ batches. Together with the adaptive-grid lower bound in Theorem 10 of the original paper, the optimal batch complexity remains $\Theta_d(\log\log T)$ when $d_z$ is unknown.

Forecasting time series accurately is critical for applications with complex data ranging from energy systems to healthcare and finance. Among current state of the art models, generative latent variable models are increasingly implemented; yet principled generalisation guarantees for modern latent variable models remain limited. In particular, while Variational AutoEncoders are widely used for sequential data, their theoretical analysis is largely restricted to i.i.d. settings. In this work, we develop a PAC-Bayesian framework for latent variables models applied to time series. Building on reconstruction-based bounds, we extend PAC-Bayesian guarantees to Markovian latent structures, capturing temporal dependencies through a sequential generative process. These guarantees do not grow with the length of the trajectory. Our bounds depend on assumptions which are common in the literature; we provide an example framework where they would be verified to show that they are not as restrictive as they may seem.

Recent hybrid Structure-from-Motion (SfM) systems combine the robustness of feed-forward 3D reconstruction with the accuracy of traditional bundle adjustment (BA) with pixel matching. They are usually the best performing methods however their scalability and usability remains limited since estimating dense correspondences between views is prohibitively costly, especially considering time constraints inherent to online applications like Visual SLAM (VSLAM). In this paper, we introduce a regularized BA framework that leverages a fast multi-view matcher and monocular priors for initialization and regularization. In contrast to existing systems, our unified approach seamlessly supports both online VSLAM and offline reconstruction from unordered image collections within the same optimization framework and sharing common hyperparameters for all tasks. Extensive experiments across both domains demonstrate improved performance and speed tradeoffs over traditional, feed-forward, and hybrid baselines. Notably for VSLAM, our uncalibrated method outperforms all previous calibrated approaches.

Dynamical symbolic regression methods identify governing differential equations from noisy data, balancing interpretability and predictive accuracy. However, standard methods often produce expressions that violate known physical laws. To address this, we propose FluxDisco, a physics-informed framework tailored for flux-based, stoichiometric ODE systems. By leveraging a known stoichiometry, we reduce the expression search space and ensure physical adherence. Our framework adapts the Monte Carlo Graph Search algorithm for the unique challenges associated with joint flux discovery of stoichiometric systems. We evaluate our method across a range of physical and biological systems, demonstrating its ability to accurately recover governing dynamics through interpretable equations.

Robotic hands vary widely in anatomical fidelity and mechanical complexity, and these structural choices influence the coordination of joint motions and the difficulty of controlling the system. A unified framework is presented in which the kinematic and actuation stages are analysed separately and in composition, through the conditioning of the task Jacobian, the actuation matrix, and their product. It is applied to two hands representing opposing design philosophies, the Shadow Dexterous Hand and the Anatomically Correct, Biomechatronic Hand, along four morphological aspects: joint axis geometry, actuator-to-DOF ratio, coupling architecture, and authority distribution. All parameters are derived from the hands' canonical digital representations. Anatomical fidelity carries no uniform advantage: oblique axes improve thumb conditioning but leave the long fingers worse conditioned than the orthogonal-axis design, while the branching tendon network improves the effective control mapping at every long finger and worsens it significantly at the thumb, where actuator authority is concentrated on thumb opposition. Predictions derived from these metrics are evaluated against reinforcement learning experiments using PPO, DDPG+HER, and TQC+HER, across three different tasks.

Multi-modal learning has demonstrated strong potential in medical applications by integrating heterogeneous data sources such as medical imaging, clinical records, and genomics to improve predictive performance and support clinical decision-making. However, advances in this area are often constrained by two key challenges: the limited availability of well-curated, ready-to-use datasets that accurately reflect real-world conditions, where medical data are frequently collected inconsistently and are often incomplete; and the inherent difficulty of integrating heterogeneous data modalities. In this work, we introduce a newly curated multi-center, multi-modal, and longitudinal dataset designed to support the evaluation of a wide range of learning pipelines under realistic conditions. The dataset comprises a total of 1,365 lung cancer patients and has three imaging modalities (whole-slide images, CT scans, and PET scans), structured clinical data, transcriptomic, and longitudinal follow-up and treatment information. For each imaging modality the dataset contains more than one instance. Moreover, the dataset exhibits substantial and non-uniform missingness across modalities, making it well-suited for studying robust multi-modal fusion strategies. We further provide both uni-modal and multi-modal benchmarks on the task of 12-month overall survival prediction, disease-specific survival, as well as longitudinal benchmark of hazard prediction under severe missing data. Our results show that, despite high levels of missingness, integrating complementary modalities consistently improves predictive performance over uni-modal approaches, highlighting the value of multi-modal fusion in realistic clinical settings. The dataset and benchmark code are available at https://github.com/ritacmendes/MMIST-LUNG.

We consider the problem of establishing limitations on the power of quantum query algorithms via the completely bounded polynomial method. In particular, we prove several optimal functional inequalities involving different notions of completely bounded polynomials. These inequalities lead to limiting theorems for the power of quantum query algorithms that improve on prior works. 1. An optimal root-influence bound for block-multilinear polynomials. Prior work showed that block-multilinear polynomials $p$ of degree $t$ satisfy a root-influence bound, $\|p\|_{\text{cb}}\geq \sum_i \sqrt{\mathrm{Inf}_i[p]}/t^2$, which is stronger than the bound appearing in the Aaronson-Ambainis conjecture. We find the optimal constant in that inequality: $\|p\|_{\text{cb}}\geq \sum_i \sqrt{\mathrm{Inf}_i[p]}/t$. Since the amplitudes of quantum algorithms that query disjoint blocks of inputs-such as $t$-fold forrelation- are block-multilinear polynomials with $\|p\|_{\text{cb}}\leq 1,$ our inequality shows that they satisfy $t\geq \sum_i\sqrt{\mathrm{Inf}_i[p]}$. We prove that this inequality yields both a more efficient classical simulation than prior results based on the Aaronson-Ambainis argument, and a qualitative improvement: all classical queries are nonadaptive. 2. Optimal Fourier growth of the highest level of quantum query algorithms. We show that for every polynomial $p$ defined on $\{-1,1\}^n$ of degree $2t$, the Fourier Growth at the level $2t,$ namely $\|\widehat p_{2t}\|_{\ell_1}$, satisfies $\|\widehat p_{2t}\|_{\ell_1}\leq (en/(2t-1))^{\frac{2t-1}{2}}\|p\|_{\text{cb}}$. This is optimal up to the factor $e$, as witnessed by $2t$-fold forrelation. As quantum query algorithms that make $t$ queries (to the whole input) satisfy $\|p\|_{\text{cb}}\leq 1$, this yields a Fourier growth bound for these algorithms, partially resolving a question by Girish (STOC, 2026).

We consider safety-critical control problems involving multiple requirements with different priorities and uncertainty in their evaluation. We represent these requirements using risk-aware rulebooks, where each requirement is assigned a risk measure and an acceptable threshold, and a priority relation is defined among the requirements. Each requirement induces a risk-evaluation function that maps a policy to the risk associated with its violation. We formulate risk-aware optimal control with rulebooks as a lexicographic optimization problem over excess risks and develop an anytime filtering and branch-and-bound algorithm that progressively tightens the certified optimality gap while characterizing the corresponding set of policies at each priority level. The algorithm returns a policy together with these gaps, which bound its suboptimality. We prove that these gaps are valid for any finite computational budget and, under additional assumptions, converge to zero as the computational budget increases. We evaluate the algorithm on a synthetic benchmark with a known optimum and a realistic highway-merging simulation with CVaR-based collision, rear-braking, headway, and comfort rules.

On-Policy Distillation (OPD) has emerged as a widely adopted post-training paradigm for enhancing large language models in reasoning domains. However, the data-centric mechanisms in OPD remain relatively underexplored. This paper presents an empirical study of data efficiency and data selection in OPD. We begin by investigating an extreme setting: training OPD on only one example, namely 1-shot OPD. Surprisingly, we find that 1-shot OPD is consistently effective across all sampled training examples and harder examples often yield superior performance gain. We next investigate what actually drives the student model's improvement in the training data. Our analysis reveals that the improvement is not driven by high token entropy, but the longer CoT paths which hard problem naturally generate. Training on longer CoT can help maintain closer alignment with the teacher over a long reasoning horizon, and learn critical thinking patterns usually missing in short CoTs, such as reflection (e.g., "Alternatively"). Based on these insights, we propose a simple data selection method that selects only hard examples for training, where even "unsolvable" examples that completely exceed the teacher's capability can be successfully used. Our experiments conducted on five models ranging from 1.5B to 8B show that training the student model on only 8 selected hard examples matches the performance of the 17K dataset baseline. Code and models will be publicly available.

The number of discrete class-separability jumps observed during ResNet finetuning is examined empirically as a predictor of final test accuracy. Across 75 experiments spanning four benchmarks (CIFAR-10, CIFAR-100, TinyImageNet, and CIFAR-10-C) and three architectures (ResNet-18, ResNet-50, and ResNet-101), with five to ten seeds per configuration, a strong within-dataset negative correlation is obtained on standard i.i.d. classification benchmarks: \(r = -0.84\) on CIFAR-10 (\(p < 10^{-8}\), \(n = 30\)) and \(r = -0.87\) on CIFAR-100 (\(p < 10^{-5}\), \(n = 15\)). Under distributional stress, the relationship attenuates: TinyImageNet yields \(r = -0.45\), and the CIFAR-10-C corruption benchmark yields \(r = -0.19\). Two additional analyses discipline the empirical claim. A partial correlation controlling for architecture depth, treated as a linear covariate, shows that on CIFAR-100 the transition count retains statistically significant predictive power (\(r_{\mathrm{partial}} = -0.69\), \(p = 0.007\)); the corresponding result under the stricter categorical conditioning is not established at \(n = 15\). A comparison against six alternative training-curve signals shows that transition count achieved the strongest correlation among the evaluated signals on CIFAR-100 and one of the strongest on CIFAR-10, but is dominated by other signals on the two stressed benchmarks. The comparison is restricted to training-curve-level signals; comparisons against effective rank, Hessian sharpness, Fisher information, margin, and neural-collapse measures, which are the strongest competitors in the current literature, are not part of the present study and remain open. The observation is presented as an in-distribution training-quality probe among a family of candidate probes, and an inexpensive detection procedure suitable for logging alongside a standard training loop is provided.

Envy-free cake cutting is a central problem in fair division with a striking divide between existence and computation. Classical topology guarantees that envy-free allocations exist, yet finding one efficiently (in the Robertson-Webb query model) has resisted decades of work. The seminal work by Aziz and Mackenzie established the existence of a bounded protocol for every $n$, but its query bound is $n^{n^{n^{n^{n^n}}}}$. Subsequently, Sokolov modified their protocol and sharpened its analysis, obtaining a bound of $n^{8n^2(1+o(1))}$ queries, the best deterministic upper bound known before this work. In contrast, the general lower bound, due to Procaccia, is merely $\Omega(n^2)$. We close much of this gap with a deterministic protocol using $n^{O(1)}2^n$ queries, the first single-exponential bound for general valuations. Like Aziz and Mackenzie, we repeatedly allocate cake without envy until some agents can leave with their pieces, guaranteed not to envy anyone who stays, even if a staying agent receives all the remaining cake. To reach this state, Aziz and Mackenzie generate $n^{n^{n^n}}$ partial allocations so that enough of them can be redistributed without conflict. Instead, we show that only polynomially many partial allocations suffice. Our key idea is to “prune” scarce allocations upfront: if only a few allocations can help a certain pair of agents, we discard them and handle that pair another way, leaving plenty of allocations for every remaining pair so each gets a distinct one. As a result, we make only polynomially many calls to Aziz and Mackenzie's subroutines, and this gives us our $n^{O(1)}2^n$ bound.

In this paper we illustrate a novel architecture generating interpretable behavior and explanations. We refer to this architecture as the Mirror Agent Model because it defines the observer model, that is the target of explicit and implicit communications, as a mirror of the agent's. With the goal of providing a general understanding of this work, we firstly show prior relevant results addressing the informative communication of agents intentions and the production of legible behavior. In the second part of the paper we furnish the architecture with novel capabilities for explanations through off-the-shelf saliency methods, followed by preliminary qualitative results.

Assessing the impacts of social policy changes is a widely acknowledged challenge for policymakers. Econometric methods can be unreliable when extrapolating to hypothetical scenarios, while field pilot programs are highly costly. In this paper, we propose using large language models (LLMs) as policy-assessment tools adapted from general-purpose models. We present FlexPension-LLM, the first domain-specialized large language model for a hierarchical pension-enrollment prediction task among flexible workers in China, and introduce DKI-RDistill, which injects policy-grounded cues into the prompt, including Probit-derived marginal effects and hukou-province pension rules. The method then uses LoRA/SFT to distill rationale-augmented supervision into an open-weight MoE student, with teacher errors corrected by regenerating those cases under ground-truth labels. On a CHFS 2019 blind split, FlexPension-LLM achieves 0.9316 Composite F1, surpassing its Claude Sonnet 4.5 teacher and 15 of 17 baselines, and is statistically indistinguishable from Claude Opus 4.6. Across four external surveys, it averages 0.7549 Composite F1 and shows the narrowest performance range among the strongest systems. Component analysis shows that gains come mainly from policy-grounded cue injection and error-filtered supervision, while rationales provide decision traces that can be checked against policy rules.

Context: Agile development methodology has been widely adopted by industry and the demand for experienced professionals in Agile-related roles is persistently high. Objectives: We focus on the learning path for a Scrum Master role in multicultural software companies and investigate the role in relation to team size, together with the learning process for a career path, and how companies monitor soft skills development. Method: We conducted our study in two phases, two qualitative surveys (interview studies) and performed a qualitative and quantitative data analysis of the results. Conclusions: Our results identified that the need for a Scrum Master (SM) depends on the size of the team, with our study indicating a six-member limit. There is no overall standardized process for soft skills learning or metrics to measure progress. Some companies measure soft skills based on feedback received from the client or from the team, and other companies are taking both types of feedback into consideration. Many learning initiatives, especially on soft skills for an SM role, were based on actions of the employees. The version of Record of this contribution is published in Software Engineering and Advanced Applications. SEAA 2025. Lecture Notes in Computer Science, vol 16083. Springer, Cham. and available online at: check DOI.

This study focuses on future Non-Terrestrial Networks (NTN) integrated with Terrestrial Networks (TN) for future 5G/6G systems. NTN envisions a 3D architecture, where Low Earth Orbit (LEO) satellite networks will play a key role in bridging the digital divide, complementing the gradual terrestrial 5G/6G rollout concentrated in high-density and high-traffic areas, by ensuring service continuity across broad geographic regions and providing coverage in case of emergencies or in remote areas. In this context, we address networking issues for the integration and federation of Terrestrial and Non-Terrestrial Network (T-NTN) in line with the IMT-2030 vision, focusing on interoperability, spectrum coexistence, unified control and management, and service continuity. Federation is a complementary approach to integration that enables distinct satellite systems to cooperate through agreements, potentially unified satellite terminals, and common resource management. We show that system federation significantly enhances both latency performance and connectivity robustness compared with non-federated LEO architectures. This paper also investigates the challenges and possible solutions for adopting the Open-RAN architecture for T-NTN, including routing options for mega-LEO systems, edge intelligence, and energy efficiency as critical elements for sustainability. Finally, we address the security, privacy, and resilience aspects of federated T-NTN architectures with emphasis on zero-trust, secure routing, trustworthy edge intelligence, Post-Quantum Cryptography (PQC), and Quantum Key Distribution (QKD).

Pathology foundation models integrated with multiple instance learning achieve competitive accuracy within single-cancer cohorts, yet cross-cancer generalization remains unresolved due to organ-specific histological and architectural differences. In this paper, we propose Conserved Immune Topology (CIT), a lightweight spatial representation for cross-cancer MSI-H prediction that augments foundation-model embeddings with biologically motivated immune descriptors. CIT uses unsupervised clustering to identify immune-associated tiles, then encodes tertiary lymphoid structures, peritumoral immune reactions, multi-scale tumor-infiltrating lymphocyte density, and immune-tumor mixing from frozen foundation-model embeddings and tile coordinates without requiring annotations or target-domain data. The proposed method was evaluated under cross-site and cross-cancer settings using CPTAC-COAD and TCGA-STAD cohorts, which introduce scanner variability, distribution shifts, and organ-specific architectural variations. Zero-shot cross-cancer transfer with CIT increased TransMIL AUC from 0.6627 to 0.7161, an absolute gain of 0.0534 (p=0.003), with consistent improvements across all three MIL aggregators. These results suggest that spatial immune topology provides potentially an organ-invariant representation for MSI-H prediction, supporting cross-cancer generalization of pathology foundation models.