Papers

Studying the dynamical behavior of a system often depends on characterizing how it transitions between long-lived states. Because such transitions are rare, observing them usually requires specialized enhanced sampling techniques. Transition Path Sampling (TPS) is a well-established method for generating reactive trajectories, which is simple to implement and does not require the definition of a preconceived reaction coordinate. However, its efficiency is limited by its sequential nature and the resulting correlations between sampled paths. Previous work addressed this limitation by combining TPS with a sampling scheme based on conditioned Boltzmann Generators, a generative machine learning model capable of sampling a given target probability distribution. This approach produces uncorrelated transition paths but relies on an accurate reaction coordinate, which is rarely known in advance. Building on recent advances in committor learning, specifically on the Artificial Intelligence for Molecular Mechanism Discovery (AIMMD) method, in this work we introduce GenAIMMD, an iterative algorithm that actively and self-consistently learns the ideal reaction coordinate (the committor) and trains a conditioned Boltzmann Generator to sample from arbitrary bias windows along it. GenAIMMD thereby provides a correlation-free and fully parallelizable path sampling scheme that does not require prior knowledge of the system's transition mechanism. We apply GenAIMMD to a two-dimensional toy model and a higher-dimensional polymer system. In both cases, GenAIMMD succeeds in training the Boltzmann Generator and learning the committor. Benchmark results show a substantial increase in performance compared to standard TPS.

While multimodal large language models (MLLMs) enable a wide range of image-text reasoning tasks, recent incidents indicate that they are vulnerable to illicit deployment and unauthorized distillation. Existing solutions for model provenance are typically confounded by shared language backbones in MLLMs and struggle to detect violations of distillation. To bridge this gap and safeguard model ownership, we present the first study on multimodal model fingerprinting. Inspired by recent findings that self-attention acts as a low-pass filter and that its low-frequency components are informative, we develop AttnPrint for white-box provenance. Specifically, we extract cross-modal attention distributions and isolate their low-frequency components to serve as model fingerprints. To facilitate black-box auditing, we further introduce DistillTrace, which employs hypothesis testing of MLLM outputs to identify potential model infringement. We conduct extensive experiments on 154 model instances across 19 multimodal architectures. Notably, AttnPrint achieves strong derivative-model detection performance while remaining robust to five downstream modification techniques. DistillTrace also provides evidence of distillation relationships under three parameter-independent techniques.

External skills provide domain knowledge and procedures without updating model parameters, but flat collections obscure skill applicability, dependencies, and composition. Graphs organize skills into addressable nodes and explicit relations, supporting selection and composition. Yet existing skill-graph methods use this structure mainly for retrieval, rather than to localize updates, scope retesting, or precisely roll back rejected changes. We introduce GRAPHSKILLAA (GraphSkill Abductive Attribution), which uses one addressable graph for skill selection, execution, failure attribution, targeted updating, validation, and rollback. Nodes separate applicability, execution, and exclusion conditions; typed edges encode prerequisite and enhancement relations. The frozen student records used nodes and edges, while the teacher contrasts related successes and failures to route each supported repair to the smallest relevant field or relation; execution lapses or insufficient evidence leave the graph unchanged. A Local Gate retests affected examples, while a Big Gate evaluates the merged graph on the complete update pool; rejected changes are rolled back. With GPT-5.6-sol, GRAPHSKILLAA reaches 81.5%, 66.7%, and 91.2% on SearchQA, LiveMath, and DocVQA, respectively, and attains the highest observed mean in every main setting. These results show that object-level attribution and graph-scoped validation make a skill graph a locally optimizable, testable, and reversible external state.

Accurate optimization of a supervised spectral objective need not produce an accurate population subspace or a better predictive representation. We investigate these distinctions for Online Kernel Supervised Principal Component Analysis (OKSPCA), which combines a centered cross-moment in finite random-feature coordinates with an Adam-style orthonormal basis update for an established objective. Fixed-map consistency, concentration and perturbation results describe the estimator and its exact subspace; same-target comparisons then assess the practical iterate separately. Across six predictive benchmarks, performance depends on the declared pipeline: replacing the tracker with the exact empirical target leaves the two regression deficits largely unchanged. Direct classification-rank models capture nearly all terminal objective energy on average, but a saved intermediate state exhibits substantial geometric deviation; a controlled sample-size study further separates empirical accuracy from population recovery. In distinct numerical-service workloads, exact on-request computation is faster in the tested classification settings, whereas Adam saves time relative to the tested full thin-SVD service for some dense wider-regression requests, alongside persistent geometric error. These diagnostics limit explanations based solely on terminal optimization accuracy and distinguish numerical cost from quality, rank coverage and freshness; they establish neither practical-tracker convergence nor predictive or deployment benefits from basis availability.

The dimension of a Galois hull is an important parameter in the construction of entanglement-assisted quantum error-correcting codes. In this paper, we study generalized Roth-Lempel (GRL) codes with prescribed Galois hull dimensions over finite fields of arbitrary characteristic. We develop a common construction method based on normalized Lagrange coefficients and suitable choices of column multipliers, and obtain six explicit constructions from multiplicative cosets, trace fibers, additive-subspace cosets, and mixed additive–multiplicative fibers. For extension sizes $s=2$ and $s=3$, we construct MDS and AMDS GRL codes with Galois hulls of specific dimensions. The resulting codes provide EAQECCs with explicit dimension, entanglement consumption, and relatively large minimum distance. Taking zero-dimensional hulls also gives Hermitian LCD GRL codes and the corresponding maximally entangled EAQECCs.

Numerical site-response predictions often deviate from observations, yet correcting these discrepancies is difficult because records are limited in both sensor coverage and number of events. This study proposes the Transfer-Enabled Forced Latent Autoencoder for Response Equations (FLARE-T) to improve these predictions by learning and calibrating low-dimensional latent dynamics that connect the base acceleration input to acceleration outputs at multiple depths. FLARE-T learns a low-dimensional response manifold and input-driven dynamics from dense finite-element simulations. It then trains a sparse encoder to map simulated sensor responses into the learned coordinates and uses limited records to calibrate the dynamics within them. A short response window initializes each prediction, while the complete base motion drives the response. The framework was evaluated using a layered-soil centrifuge test and the Lotung field vertical array. Test-set results show that FLARE-T improved multi-depth acceleration histories and 5%-damped pseudoacceleration response spectra relative to the original finite-element models, reducing errors at every evaluated sensor for motions of different intensities and, at Lotung, for both horizontal components. Two Lotung source models with different constitutive parameters achieved comparable test-set accuracy, indicating reduced dependence on precise prior calibration. FLARE-T therefore provides a data-efficient means of combining dense numerical response information with limited field records to improve future site-response predictions.

This paper studies resource-limited external influence in time-varying opinion networks when a controller must choose both a small set of agents and a scalar intervention at each update. We use the affine free response, which includes DeGroot and Friedkin–Johnsen dynamics, followed by a direct sparse action. Eliminating the scalar action reduces the one-step problem to a cardinality-constrained support objective whose global optimum is obtained from the largest or smallest residual components. Hence exact sparse one-step-ahead optimal control requires one sort and \(O(|\mathcal T|\log|\mathcal T|)\) operations, where \(\mathcal T\) is the support set. For periodic affine dynamics, the resulting state-dependent low-rank feedback admits cycle-based ultimate bounds and exact tracking of model-invariant targets. The DeGroot and Friedkin–Johnsen specializations expose the role of target invariance and the affine mismatch term. With several competing external players, each sparse best response retains the same sorting structure and the complete binary–continuous stage game is an exact potential game, so a pure-strategy equilibrium exists at every frozen state. Same-state welfare benchmarks separate loss due to scalar competition from the additional loss due to strategic support selection. Numerical examples validate exact support selection and illustrate switching-order effects, DeGroot/Friedkin–Johnsen tracking, resource tradeoffs, and competitive implementations.

The economic value of inference depends on how capacity and task information are distributed across stages of AI production. We study these organizational margins using controlled workflow experiments on externally verified software-engineering tasks. In two matched resource panels, direct execution records the same success rate of 59.6 percent at logical-token ceilings of 12,000 and 24,000, while success under information-constrained planning rises from 36.2 to 51.2 percent. The planning disadvantage narrows by 15.0 percentage points (95 percent task-cluster bootstrap interval: 4.2 to 25.8). A strict read-only planning campaign varies whether the planner sees the task issue. At 12,000 tokens, issue access raises success by about 16 percentage points over issue-hidden planning. Compared with direct execution, task-informed planning is about 10 points lower at 12,000 tokens; at 24,000 tokens, it shows a 29.6-point advantage. In the resource panels, direct execution uses substantially less than either ceiling, while the planning workflow's binding rate falls from 46.2 to 0.8 percent and downstream execution accounts for 89.9 percent of the increase in total use. Scale determines the capacity available to a system; workflow and information structure shape the productive value

Coding agents and evolutionary code-search systems can improve implementations once a target and evaluation criterion have been specified. Applying these methods to an existing software repository raises an earlier question: which implementation choices are worth investigating for a high-level engineering objective? We introduce optimization-opportunity discovery, the repository-level task of identifying candidate source regions, explaining how they relate to the objective, and proposing possible changes. The task takes as input a repository, an engineering objective, and optional runtime evidence such as offline telemetry observations or profiles. It does not require the user to specify a defect, bottleneck, or code location. We present Spotlights, a system that performs this task through logical repository mapping, successive agent reviews, and optional research linking candidates to relevant techniques. We evaluate Spotlights across model serving, document retrieval, blockchain ordering, and document processing. Across three cases, it recovers seven of nine expert-selected targets. In the reliability study, 70% of the top ten candidates meet the stated correctness and severity thresholds. Across five repeated retrieval runs, 73.6% of candidate occurrences have a matching source region in all five runs. Spotlights also rediscovers the target of a withheld retrieval optimization and connects it to a relevant tiling technique. In an implementation study, a discovered change reduces end-to-end page-processing runtime by 10.6% while preserving measured output quality. These results establish optimization-opportunity discovery as a distinct and empirically evaluable step between a broad engineering objective and subsequent implementation and validation.

We design and analyse a hybrid high-order (HHO) method for the linear elasticity problem with a linear slip interface, i.e.\ with interface conditions of spring type, in which the jump of the displacement across the interface is proportional to the traction through a compliancy tensor $\bK=\alpha\bI+(\beta-\alpha)\udl{n}\otimes\udl{n}$. The mesh is fitted to the interface, the discrete unknowns are polynomials of degree $k\ge 1$ on the mesh faces and of degree $k+1$ in the mesh cells, and general polytopal cells are allowed. The two novelties of the method lie in the local symmetric strain reconstruction, which incorporates the interface condition, and in the interface stabilisation, which is built from a regularised interface stiffness $\bS_h$ in the spirit of Hansbo and Hansbo \cite{HH04}. As a consequence, a single formulation covers the whole range of compliancies, from the perfectly bonded interface $\alpha=\beta=0$, where the interface stabilisation acts as a Nitsche-type penalty, to the traction-free interface obtained as $\alpha,\beta\to+\infty$, and no unknown is attached to the jump. We prove that the discrete bilinear form is coercive and that the errors converge as $h^{k+1}$ in the energy norm and as $h^{k+2}$ in the $L^2$ norm, with constants that are independent of the compliancy parameters and of the Lamé coefficient $\lambda$, so that the method is also locking free. Various numerical examples and comparisons are provided to confirm the theoretical results.

Robots searching for a target from a natural-language description must determine not only where to search, but also which observed candidate is the desired target. These decisions reflect two distinct sources of uncertainty - spatial uncertainty over candidate locations and semantic uncertainty over target identity - that are often conflated in VLM-based search systems. We introduce a spatial-semantic uncertainty formulation that maintains separate beliefs over each component and integrates probabilistic VLM evidence into a global target-identity posterior, including probability mass for undiscovered targets. This decomposition allows an information-theoretic planner to independently value candidate discovery and target disambiguation through spatial and semantic expected information gain (EIG), providing an explicit mechanism for trading broader exploration against earlier identification. We evaluate six VLM uncertainty-elicitation interfaces on 500 synthetic targets and show that similar recognition accuracy can conceal substantial differences in calibration and false confidence. In degraded-observation search-and-identify experiments, EIG-based planners reach confident decisions in 75.0%-92.5% of trials, compared with 20.0% for Random search, while different spatial-semantic weightings achieve comparable identification accuracy once confidence is attained. Increasing semantic emphasis reduces unnecessary exploration and VLM queries, demonstrating that explicitly planning over semantic uncertainty can accelerate target resolution without sacrificing decision quality. These results highlight the distinct roles of uncertainty representation and uncertainty-driven planning in embodied VLM systems.

Geospatial foundation models can provide strong flood-segmentation performance, but their size limits deployment on memory-constrained edge hardware. We distill a 300-million-parameter Prithvi-EO-2.0 teacher, fine-tuned on the 252 manually labeled Sen1Floods11 training scenes, into a 0.7-million-parameter EfficientViT-B0 student. The teacher supervises additional unlabeled Sentinel-2 imagery, allowing the student training set to grow without new manual annotations. At the matched budget of 252 scenes, teacher-supervised training is competitive with direct training and improves STURM-Flood performance across tested configurations; a geometry-matched control shows that label source alone does not explain the difference. Scaling the teacher-supervised pool to 2,500 scenes narrows the remaining student–teacher gap: the float student reaches 0.787 water intersection over union on the Sen1Floods11 test split against 0.822 for the teacher, matches the teacher on STURM-Flood under our evaluation protocol, and remains below it on WorldFloods-v2. After activation replacement and quantization-aware training, the student runs as a 1.5-megabyte 8-bit integer (INT8) TensorRT engine on a Jetson Xavier NX at 5.57 milliseconds of graphics processing unit (GPU) compute per 512-by-512 image, with approximately 14 megabytes of runtime device memory. A fixed modified normalized difference water index (MNDWI) threshold is competitive with both models on the two clean external benchmarks, so we interpret those benchmarks as generalization tests rather than as evidence of learned-model superiority over a spectral rule. The results support the conclusion: foundation-model supervision can amplify a fixed manual annotation budget into a substantially larger training set and yield a compact, deployable edge model.

Validating variational quantum eigensolver (VQE) pipelines on realistic chemical systems is a critical stepping stone toward tackling classically intractable molecules. In this work, we report emulator-based validation of the VQE workflow proposed for the copper active site, executed on the LRZ Eviden Qaptiva emulator. A problem-tailored ADAPT-VQE with a full singles-doubles-triples (S+D+T) pool removes over 99% of the correlation error, reaching a stable plateau at 4.03 mEh relative to exact diagonalisation (148 selected operators). This establishes a >40x improvement over the hardware-efficient ansatz (HEA). Finally, we frame these emulator results as a robust baseline for upcoming physical quantum hardware runs on Euro-Q-Exa.

This article presents a mathematical model of the Motivated Emotional Mind cognitive architecture developed for embodied intelligent systems. Such a system learns to maintain its homeostasis through a generalized form of reinforcement learning based on its internal motivations, termed motivated learning (ML). The principal contribution of this article is a rigorous formalization of the re-entrant loop integrating feedforward processing, lateral interactions, and feedback pathways, together with the representational selection mechanisms that govern adaptive system responses. The model specifies how ongoing exteroceptive and interoceptive signals, bodily-motivational context, and memory traces are bound into associative memory structures termed semblions, which compete for access to further processing and top-down reconstruction. The formalization encompasses secondary perception, representational competition, curiosity, procedural gaps, and action selection directed toward limiting allostatic violations. Within this framework, motivated learning is tailored to embodied systems whose dynamics are shaped by needs, affect, and the current regulatory state. Unlike standard reinforcement-learning models, the proposed approach incorporates need thresholds, goal generation and shifting goals, bodily state, resource constraints, and action uncertainty, thereby providing a more adequate account of response selection under regulatory pressure. Global affect functions as a central control signal, modulating the learning rate, representational valence, and the balance between exploration and exploitation. The model presented here is a step toward a more rigorous formalization of cognitive phenomena and may provide a basis for further theoretical analysis, computer simulation, and implementation in artificial-intelligence systems inspired by biological processes.

Autonomous mobile robots require timeefficient planning and safety-critical dynamic obstacle avoidance under constrained onboard computation. While Iterative Learning Planning (ILP) offers lightweight and efficient traversal planning, it lacks explicit mechanisms for dynamic obstacle perception and avoidance. This article extends ILP to safety-critical navigation in dynamic environments by integrating an anticipatory risk-blended control barrier function (ARB-CBF). The extended ILP learns traversal-speed and steering-bias profiles via a fractionalpower update based on local obstacle risk, generating nominal control commands that ARB-CBF modifies at runtime for real-time safety guarantees. Algorithmic analysis demonstrates that the ILP replanning stage scales at O(kN) for k iterations and N waypoints, while ARB-CBF executes with linear complexity. Comprehensive simulations and real-world experiments validate the framework, demonstrating superior temporal efficiency and safety with lower computational overhead compared to optimizationbased baselines, making it highly suitable for resourceconstrained platforms.

Automated pain recognition from facial expression could make continuous welfare assessment practical in sheep, but adoption depends on trust: a stockperson cannot act on a score that arrives without justification. We ground a model in the Sheep Pain Facial Expression Scale (SPFES) by letting each detected facial region attend over text embeddings of the clinical descriptors and then test whether the resulting explanations mean anything. They do not. Ablating an entire descriptor changes the predicted logit by about $10^{-4}$, and the most-attended cue agrees with the predicted pain level in only $32.6\%$ of regions, although the attention maps, the learned gate, and the generated text all proposed otherwise. We therefore remove the appearance bypass with a concept bottleneck whose classifier reads only SPFES concept scores, supervised by per-region state annotations that image-level pipelines discard. This costs $0.05$–$0.10$ in Cohen's $\kappa$ but yields concepts that are demonstrably learned: minority pain-indicating states are recovered at $3.5$–$8.3\times$ their base rates, and the ear and eye severity orderings emerge without severity supervision. Removing the supervision alone leaves $\kappa$ unchanged while concept accuracy falls to $0.109$, showing that architectural necessity does not imply semantic validity. We also show that pooled concept accuracy is misleading under clinical imbalance and provide a cross-validated, protocol-matched benchmark of seven methods on this dataset.

We study income taxation when an AI agent implements economically relevant choices through a rule hidden from the government. Alongside unobserved productive ability, this hidden preference-to-execution mapping creates double unobservability: the same observable tax-base response can carry different welfare consequences. We call the resulting income welfare-opaque. Our constructions show that tax-base statistics can coincide while reform welfare effects differ, even when mechanical welfare weights are identical. We derive an optimal-tax condition that adds a response-weighted execution wedge to the familiar sufficient statistics. A higher marginal rate gains a corrective benefit under local over-execution and an additional cost under local under-execution. Observing the wedge identifies the welfare effect of a marginal reform at the prevailing schedule; bounds on it deliver bounds on that effect. A controlled laboratory compares 4,500 model runs across five AI engines. Faithful delegation selects the score maximizer in essentially all runs. Conflicted objectives produce heterogeneous responses: Claude largely preserves the score maximizer, GLM moves predominantly downward, and GPT-mini and Qwen show concentrated lower-tail increases. Qwen also makes substantial downward adjustments. Different engines locate their departures at different points and in different directions of the designed distribution. Explicit scores align model rankings; formula-based objective instructions yield more uneven agreement. Qwen shows a clear positive tax-by-objective interaction, but its direction does not generalize across engines and the pooled sign depends on its inclusion. The analysis identifies execution information as a complement to conventional tax-base statistics.

Document parsing converts document images into structured content and requires reliable performance across diverse layouts and acquisition conditions. Yet training corpora are biased toward common document types and clean digital pages, while expanding coverage alone does not specify how to address a parser's remaining weaknesses. We present WeVisDoc, a two-stage data-centric framework for robust end-to-end document parsing. Stage I broadens semantic, structural, and appearance coverage through heterogeneous data and structure-preserving degradation synthesis. Stage II uses a held-out probe to measure the Stage I parser's residual errors within fixed visual-structural clusters. These diagnostics guide targeted data construction and reallocation of the target-token budget. WeVisDoc-4B achieves an Overall score of 95.38 on OmniDocBench v1.6 and a mean Overall score of 75.54 across the three PureDocBench tracks, ranking first among the compared end-to-end parsers in all four settings. Compared with Stage I, Stage II improves Overall scores for the 2B and 4B models on both benchmarks, with larger gains on the degraded PureDocBench tracks, including a 4.03-point gain for the 4B model on the Real Degraded track.

Graph foundation models (GFMs) aim to learn transferable representations across severely heterogeneous graph domains. However, severe domain shifts in topology, graph scale, and feature semantics impede the construction of a unified, domain-agnostic representation space. To address this, we propose SCGFM-ART, a structure-centric GFM framework that aligns arbitrary graphs onto a shared relational atlas via Amortized Relational Transport (ART). The relational atlas serves as a universal coordinate system defined by a finite set of relational landmarks (bases), while ART directly predicts reusable, end-to-end graph-to-base transport plans, bypassing costly runtime Gromov-Wasserstein optimizations. Under this formulation, SCGFM-ART decomposes a graph into a unified representation: globally via its relational response coordinates relative to the atlas, and locally via its node-to-role structural correspondences. These correspondences project disparate node attributes into a canonical role space, resolving structural and semantic heterogeneity within a singular alignment interface. Rigorously modeling graphs and atlas bases as finite measured relational spaces, we establish coordinate fidelity bounds, prove stability under predicted transport plans, and derive an amortized coverage bound that guarantees our learning objective tightly surrogates ideal relational coverage. Benchmarked across 14 cross-domain graph- and node-level classification tasks, SCGFM-ART achieves state-of-the-art transferability, securing superior average ranks of 2.29 and 1.14, respectively. Topological perturbation analyses demonstrate that node-role transport retains fine-grained structural nuances beyond global coordinates. On real-world benchmarks, the amortized formulation yields 44.2 to 85.1 times faster frozen target-domain inference by avoiding iterative alignment at test time.

We develop a new framework for proving the undecidability of first-order theories of structures of the form $\langle \mathbb{N}; +, P \rangle$, $\langle \mathbb{N}; <, f \rangle$, and $\langle \mathbb{N}; +, f\rangle$, where $P \subseteq \mathbb{N}$ and $f \colon \mathbb{N} \to \mathbb{N}$. It is based on the recent proof of Hieronymi and Schulz that the first-order theory of $\langle \mathbb{N}; +, \{2^n \colon n \in \mathbb{N}\}, \{3^n \colon n \in \mathbb{N}\}\rangle$ is undecidable, and capable of transforming various randomness results about integer sequences into undecidability proofs. We apply our method to a large class of integer linear recurrence sequences, as well as various special functions, in particular showing that the first-order theories of $\langle \mathbb{N}; +, \{u_n \colon n \in \mathbb{N}\} \cap \mathbb{N}\rangle$, $\langle\mathbb{N}; <, n \mapsto \max\{0,u_n\}\rangle$, and $\langle \mathbb{N}; <, \phi\rangle$ are undecidable, where $(u_n)_{n\in\mathbb{N}}$ is any integer LRS with exactly two non-repeated dominant roots satisfying a non-degeneracy assumption, and $\phi$ is Euler's totient function.

Tactile signals provide direct contact and force measurements that are essential for understanding physical interactions and enabling dexterous robotic manipulation. However, tactile sensing requires direct measurement at contact interfaces, making large-scale data collection reliant on intrusive, costly, and restrictive instrumentation. We present TouchSight, a monocular egocentric vision framework for dense full-hand contact force prediction that leverages 500 hours of pressure-glove recordings and extensive hand-object interaction (HOI) data. To address the appearance gap between gloved training data and bare-hand real-world scenarios, we construct TwinTouch-20H: 20 hours of paired visual data in which generative video models re-render gloved recordings as bare-hand observations against new backgrounds while preserving the original measured tactile labels. TouchSight predicts dense force from both gloved and generated bare-hand videos, outperforms prior contact prediction methods on OakInk2, qualitatively generalizes to natural bare-hand egocentric videos from unseen datasets, and improves consistently as glove supervision scales. These results demonstrate that dense tactile signals can be recovered from egocentric vision alone, without tactile instrumentation at capture time.

When aligning frontier models through post-training techniques, it is not possible to directly demonstrate all of the behaviours we want a model to exhibit in all possible deployment environments; our model must generalise outside of the post-training distribution. One proposed solution is alignment midtraining (AMT), which continues pretraining on large volumes of alignment-relevant documents to encourage generalisation in later stages of training. Despite the prominence of AMT as an alignment approach, there is limited public evidence for its effectiveness. To resolve this, we identify several assumptions around midtraining and evaluate them across scale: up to 110 billion-parameter models and 1 billion midtraining tokens. For instance, we study a scenario where post-training data is ambiguous between two possible motivations. We find that midtraining can steer the model's motivation in simple versions of this setting. However, the presence of a tiny fraction of finetuning data which suggests a competing motivation erases the effects of AMT. We also study scenarios in which we want an AI to follow a number of rules, but only demonstrate a subset of them. We find that demonstrations must be present either in midtraining or post-training datasets for these rules to be robustly learned. Based on these and other findings, we do not believe that there is sufficient public evidence for us to confidently state that midtraining can address the core difficulties inherent in aligning powerful AI systems.

Background: Code quality metrics are intended to measure latent properties of software source code. Although numerous code metrics have been proposed and used, their construct validity is rarely evaluated. Thus, the extent to which code metrics actually measure what they claim to measure is often unclear. Aim: Drawing from modern measurement theory, we investigate the construct validity of common class-level, object-oriented code quality metrics by identifying their factor structure using Exploratory Factor Analysis (EFA). The metrics were extracted from the Apache Maven project by three software tools: Designite, JHawk, and Understand. The factor structure was later verified using Confirmatory Factor Analysis (CFA) on 22 randomly selected open source projects meeting a predetermined eligibility criteria. Results: 24 code quality metrics that correspond to six constructs: Cohesion, In-Coupling, Out-Coupling, Size, Sub-Inheritance (related to subclasses), and Sup-Inheritance (related to superclasses) were revealed in the underlying factor structure. Ten metrics did not correspond to any known dimension of software quality and were removed in the EFA. Ten additional metrics exhibited low loadings in the CFA, suggesting their removal from the final measurement model. Size, Cohesion, Inheritance, and Coupling were the constructs retained, with subcategories identified for Inheritance and Coupling. Conclusions: Our results strongly support the construct validity of 24 code quality metrics. Coupling and Inheritance are revealed as multidimensional constructs, since they require measuring two different concepts, revealed as sub-categories in our analysis, and Complexity may be better explored in a multilevel model. Overall, our study demonstrates the value of applying modern measurement theory and latent variable modeling in validating software code quality metrics.

In the first half of the paper we revisit the question regarding MaxEnt probability distributions on spheres. We derive families of MaxEnt distributions on spheres in real and complex vector spaces with fixed expected Busemann function (energy). As a particular case, we deduce sub-families on canonical energy levels where inverse temperature equals the dimension of the sphere. In the second part we focus on the information manifold of canonical MaxEnt distributions on the sphere in the complex vector space. This manifold is isomorphic to the Bergman ball. We introduce the reproducing kernel on this manifold and use the RKHS theory to elaborate the model of cognition. In particular, we state the principle of minimal cognitive effort in RKHS and demonstrate its dual relationship with the MaxEnt principle for probability distributions on the boundary sphere.

Two-timescale stochastic approximation (TTSA) is a fundamental tool for analyzing coupled iterative algorithms in reinforcement learning, optimization, and stochastic control. However, finite-time guarantees for nonlinear two-timescale schemes remain difficult to obtain, especially under constant step-sizes. In this paper, we study nonlinear TTSA with step-sizes $\alpha\gg\beta$. Under standard stability, regularity, and Markovian noise assumptions, we upper bound the mean-squared error and the bias of both iterates around their limiting equilibria. Our bounds scale as $O(\alpha+\beta^2/\alpha^2)$, which we prove to be tight when $\beta\le\alpha^{3/2}$. The analysis separates the contributions of initial conditions, fast-timescale tracking error, Markovian dependence, and timescale coupling, thereby clarifying the origin of the $\beta^2/\alpha^2$ term. Our results reveal qualitative differences from the linear TTSA setting previously studied, showing that nonlinear dynamics introduce additional finite-time effects that are absent in the linear case.

Large language models are usually interpreted through concepts that humans already possess: truthfulness, refusal, deception, personality, harmfulness, and related categories. This paper asks whether models may also represent and use distinctions for which no adequate human concept exists. We call such internal structures xeno-representations, and their study xeno-interpretability. We distinguish the human-interpretable semantic space from the xeno-semantic space: the region of model-native representations for which no adequate human conceptual counterpart is available. We show that the space of possible internal distinctions in an LLM is substantially larger than the space available through finite human descriptions. We then separate experimental identification from semantic interpretation: an internal representation may be reproducibly located, geometrically characterized, causally manipulated, and linked to downstream behaviour even when its semantic content cannot be adequately expressed in human terms. On this basis, we sketch an empirical programme to identify xeno-representations. We finally examine the implications for AI safety and multi-agent systems, where model-native representations may propagate and stabilize across interacting agents while remaining only partially visible through human-readable communication. Xeno-interpretability therefore shifts the aim of interpretability from finding human concepts inside models toward discovering and characterizing the representational structures that are native to the models themselves and might affect their behaviour in unpredictable ways.

Free-flying robots rely on multiple thrusters to maneuver in space. If one or more of these thrusters fail, the robot may lose control authority and risk mission failure. At the same time, their free-flying nature implies that, even in the absence of actuation, they continue along (locally) straight-line trajectories. In this work we present a probabilistic, proactive, motion planning framework that explicitly accounts for actuator failures in space. We model actuator failure modes as a Markov chain and propagate the probability of successfully reaching the goal along the planning horizon. Precomputed reachable sets evaluate the robot's capabilities of reaching waypoints under potential failures and an RRT$^*$-based planner concatenates these waypoints. The resulting algorithm maximizes the overall target-reaching probability, providing maximally resilient motion plans utilizing free-flying properties. We validate our approach experimentally on a physical free-flyer platform with injected actuator failures.

Agricultural soils are a major untapped carbon sink. Carbon farming is emerging as a promising practice for tapping this potential. Smallholder farmers, who dominate agriculture across South Asia and sub-Saharan Africa, are key to scaling climate mitigation via carbon farming. It is ironic that real-world carbon programs largely fail to reach them. We study this important gap through the lens of contract design. An aggregator offers a single pooled contract to a heterogeneous population of smallholder farmers who have private adoption costs (adverse selection) and exert unobserved effort (moral hazard), with agronomic outcomes evolving over multiple seasons. We formulate this evolving contracting problem as a POMDP and use reinforcement learning to learn a dynamic profit-maximising contract. We analyse the performance of the aggregator under various conditions. We find that a profit-maximising aggregator does not merely inherit the exclusion of smallholders, it amplifies it. On large farms the aggregator realises 87.7% of achievable adoption, against only 8.2% on smallholdings. Per-hectare Measurement, Reporting and Verification (MRV) costs fall as farm size rises, and the aggregator's pooling contract compounds this gradient rather than offsetting it. A counterfactual that makes MRV costs purely area-proportional eliminates this disparity. Our results and simulation can guide contract and policy design that opens carbon income to smallholders while enabling agricultural soils to contribute to climate mitigation at scale.

Let $q=3^m$, let $K=\mathbb F_{q^2}$, and let \[\mathcal T=\{x\in K^*:\operatorname{N}_{K/\mathbb F_q}(x)=1\}.\] For both cyclic and constacyclic Gashkov-Sidel'nikov codes, we show that the set of signed parity-check column labels is precisely $\mathcal T$. Consequently, the decoding problem separates into two stages: determining the minimum error weight associated with a syndrome $S$ and constructing an error vector attaining this minimum. We identify the former quantity with the minimum additive length of $S$ with respect to $\mathcal T$ and determine it exactly by the norm and the quadratic character of $\mathbb F_q$. We also determine the complete coset-weight distribution and recover the known covering radius $3$. For the constructive part, we use quadratic-character sums and Weil bounds to construct a coset leader for every syndrome of coset weight three. The resulting procedures give complete maximum-likelihood decoders.

Semantic parsing (SP)-based knowledge base question answering aims to answer natural language questions by generating executable logical forms (LFs) over knowledge bases (KBs). When applying Large Language Models (LLMs) to this task, a key challenge over large, heterogeneous KBs is selecting question-related schema elements (i.e., relations and classes) and composing them into complex LFs. Recent LLM-based methods often make early discrete commitments to schema elements during intermediate reasoning, allowing incorrect intermediate schema decisions to propagate and finally result in incorrect LFs. To overcome this limitation, we propose SALR, a schema-anchored latent reasoning method for LF construction. It performs multi-step reasoning by generating continuous thoughts in the model's hidden states, thereby delaying the explicit commitment to LF decisions. To ground this latent reasoning process in the corresponding KB schema, SALR aligns continuous thoughts with a codebook of KB schema elements through an alignment objective supervised by schema traces deterministically derived from gold LFs. It then incorporates the aligned schema codes into inputs for subsequent reasoning steps. This schema-mediated feedback guides LF generation without requiring the model to emit an explicit textual reasoning trajectory. Experiments on GrailQA and WebQSP show that SALR achieves consistent overall gains over strong baselines. Notably, on compositional questions from GrailQA, SALR outperforms TIARA, a strong SP-based baseline, by 2.86 F1 points. Further analyses show that schema-mediated feedback affects LF generation and that schema information is recoverable from the latent states.