Papers

Training LLM agents to interact with users over multiple turns is challenging because the learning signal depends not only on the agent, but also on the user simulator and environment, which jointly shape the agent's experience. Simulated users can be stochastic or inconsistent, while environments may contain unreliable tools, incomplete state coverage, and incorrect verifiers. Prior work mainly addresses the resulting instability through reward shaping or denser supervision, while treating these components as fixed. We instead introduce EAU, a meta-agent framework for environment–agent–user co-evolution. During training, EAU monitors trajectories, tool traces, rewards, and training statistics; diagnoses failures and uninformative interactions; and proposes targeted modifications to the user simulator, environment, reward design, and RL configuration. Validated changes are hot-swapped into the ongoing run, allowing the training setup to evolve together with the policy. These adaptations also persist across runs, so later training can build on previously discovered improvements rather than restarting from a fixed setup. We enable this process through an editable symbolic representation of user and environment dynamics, while the agent remains a neural LLM optimized with RL. Across three interactive-agent benchmarks and two backbone models, EAU consistently outperforms prior RL methods, improving the overall score over the strongest baseline by 14.7–15.0% and increasing successful task completion on ColBench by up to 36.7%, showing that co-evolving the training system with the policy can substantially improve multi-turn RL.

Safe robot navigation in unknown, dynamic environments requires more than geometric collision checking: a robot still far from an obstacle may be unable to avoid it because of its speed, orientation, or the obstacle's motion. Hamilton–Jacobi (HJ) reachability analysis captures this in a value function whose sign separates states that admit a safe continuation from those where collision is unavoidable, but computing it online for every newly observed environment is too expensive. This dissertation trains neural implicit representations offline on reachability solutions to give optimization-based controllers this information in real time. Since the value function never exceeds the constraint, it equals a directly computed constraint, such as a signed distance, minus a nonnegative dynamics correction. Learning only this correction keeps represented obstacles outside the predicted safe set despite approximation errors. The first two of five methods use the learned function as a terminal constraint in model predictive control, for static local environments and predicted obstacle motion. The next two use it as a barrier candidate in control barrier function safety filters, for static environments and, through pairwise barriers in relative coordinates, for a varying number of moving obstacles. Simulations and hardware experiments on ground and aerial robots show better collision avoidance than geometric and hand-designed baselines, especially for short horizons and moving obstacles. The fifth method replaces the grid-based offline solver, which limits the first four to low-dimensional models. For control-affine dynamics with polytopic inputs, the HJ Hamiltonian is maximized at a vertex, so a parallelizable tree search over vertex inputs can generate training labels. With a loss from the HJ variational inequality, these labels yield barrier candidates for systems with up to 13 states.

Neural networks (NNs) have been employed in various ways to accelerate the solution of multi-parametric optimization problems (OPs), which prominently arise in model predictive control (MPC). Early approaches focused on approximating the parametric solution (e.g., the optimal control law) directly by NNs. More recently, learning-based techniques have been used to emulate or enhance solver iterations. For the direct approximation of the parametric solution, it has been demonstrated that exploiting structural properties of the underlying OP facilitates the design of suitable NN architectures. In this paper, we show that structural analysis can also be leveraged to systematically integrate learning into classical optimization algorithms, thereby enhancing their performance. We formalize our concept for the case of multi-parametric quadratic programs (mpQPs), which frequently arise in linear-quadratic MPC, as a representative application. More precisely, we demonstrate that tailored NNs enable a parameter-dependent fine-tuning of the step sizes in a standard dual projected gradient scheme. The design of the NNs is guided by the observation that optimal step sizes, which permit obtaining the exact optimal solution in finitely many iterations, are piecewise constant functions over the parameter space (e.g., the state space in MPC). We establish key theoretical properties of the proposed approach and illustrate its effectiveness in a numerical case study.

We study the last iterate of stochastic gradient descent (SGD) on streaming least squares with discretely distributed features (Slock's model), which decouples the dynamics across eigendirections. Under power-law source and capacity conditions, we derive the exact finite-time risk and asymptotic equivalents of constant and Warmup-Stable-Decay (WSD) schedules. With a tuned base learning-rate, the resulting scaling law decays as $\Theta(T^{-q})$, where $T$ is the horizon, and $q$ is independent of the cooldown length $c \in (0,1]$ but is strictly smaller for the constant schedule: any cooldown of fixed relative length $c > 0$, however short, improves the risk by a polynomial factor, while its length only affects the constant, which is minimized by linear decay ($c = 1$). Finally, we show that the optimal risk depends on the batch size $b$ and horizon $T$ only through $bT$, up to a critical batch size.

Vision-Language-Action (VLA) models are often pretrained via imitation learning, making reinforcement learning (RL) post-training important when deployment moves beyond demonstration coverage. Existing refinement paradigms expose a reliability–efficiency tension: model-free refinement avoids simulator bias but is interaction-intensive, whereas model-based refinement can exploit unreliable imagination even as its simulator adapts. We propose Model-Evolving Reinforcement Learning (MERL), a trust-calibrated framework that couples policy optimization with an online-updated simulator comprising a world model and a reward proxy head. Across stages, simulator reliability controls the imagined-data ratio and rollout horizon; within rollouts, chunk-level trust controls replay selection and optimization weight. For recursive imagination without exact grounded continuations, a residual predictor estimates trust without querying future ground truth. On perturbed LIBERO-PRO tasks, MERL reaches 79.7% average success, improves high-threshold learning speed by 10.9% over the matched model-free control, and outperforms same-protocol external baselines across three training seeds. Fixed-data adaptation from 50 demonstrations per task achieves 80.0% average success on two physical-robot tasks.

In parallel decoding, diffusion language models commit several tokens per forward pass, so co-committed tokens cannot see each other, breaking multi-token format units that parsers need. On ECHO, an open-weight one-step chest X-ray (CXR) report generator, 32.7-36.7% of evaluation reports lack a parseable Impression heading and score as empty. We propose intent-triggered anchoring, a training-free decoding rule that completes a format unit once the model commits its first token, inserting only units with a unique, content-free completion; fallbacks write the unit when the model stops or loops without starting it. On three CXR evaluation sets, the rule raises readable rates from 63.3-67.3% to 100.0% with no regressions and a 95th-percentile latency ratio of 0.68-1.28 against official decoding. Broken headings state findings absent from the reference; adding the heading at the end leaves them in 15.5-31.0% of reports, our rule in at most 0.6%. For math answer markers, the rule raises accuracy in the most parallel setting of five diffusion language models by 5.8-20.8 points on GSM8K and 5.1-17.1 on GSM-Symbolic, held out from tuning. With lenient answer extraction, gains are 3.3-12.5 and 2.7-6.3 points; the rule adds 1.4-8.6 points over a stop-only fallback for four of the five models.

Pattern mining is a valuable tool for exploratory data analysis, but identifying relevant patterns for a specific user is challenging. Various interestingness mea- sures have been developed to evaluate patterns, but they may not efficiently estimate user-specific functions. Learning user-specific functions by ranking patterns has been proposed, but this requires significant time and training samples. In this paper, we present a solution that formulates the problem of learning pattern ranking functions as a multicriteria decision making problem. Our approach uses Analytic Hierarchy Process (AHP) to elicit weights for different interestingness measures based on user preference. We also propose an active learning mode with a sensitivity-based heuristic to minimize user ranking queries while still providing high-quality results. Experiments show that our approach significantly reduces running time and returns precise pattern ranking while being robust to user error, compared to state-of-the-art approaches.

Developing reliable predictive models for multi-physics systems often forces a compromise between computationally demanding physical simulators and data-hungry, opaque machine learning (ML) algorithms. While incorporating dimensional analysis into directed computation graphs offers a structurally sound alternative to explicitly solving partial differential equations (PDEs), such formulations typically yield underdetermined systems that act as imperfect simulators. We introduce pi-Net, a physics-informed latent variable model that resolves these imperfect simulators. The hybrid pi-Net architecture consists of an encoder that maps inputs into physically normalized latent space, which are then decoded into final predictions by a differentiable physics head. Evaluated on a real-world, data-scarce Gas Metal Arc Welding (GMAW) case study, the framework achieves an R2 of 0.864 with only 60 experimental samples. The results demonstrate that pi-Net significantly outperforms purely data-driven baselines in data-scarce regimes while providing causal, process-level explainability.

Discovering and disambiguating entities from multimodal and heterogeneous data is a key challenge in discovery science, where relevant knowledge is often implicit and distributed across modalities. We present KARAMEL, a framework for multimodal entity linking (MEL) - the task of aligning visual and textual mentions to entities in a knowledge base - in the cultural heritage domain. Unlike prior vision-language models that primarily rely on surface-level multimodal alignment, our approach incorporates sociohistorical and contextual knowledge through a knowledge-aware architecture that fuses a vision-language encoder with a heterogeneous graph neural network. This design enables contextual reasoning over complex relational structures and supports robust entity disambiguation under ambiguous or underspecified evidence. To benchmark multimodal discovery in this setting, we introduce WikiMuSA (Wikidata-based Multimodal Semantic data for Art), a new dataset that links artworks to Wikidata entities via images, textual descriptions, and structured contextual knowledge. Experimental results show that KARAMEL surpasses strong baselines, particularly on implicit or abstract mentions, while ablation studies highlight the critical role of domain-specific, structured context in learning expressive representations. While instantiated in the cultural heritage domain, KARAMEL addresses a general knowledge discovery problem and applies to other scientific settings involving multimodal observations and heterogeneous background knowledge. Both the dataset and source code are made publicly available.

Terrestrial network (TN) failures during disasters can overload surviving base stations (BSs) and disrupt emergency services. A non-terrestrial network (NTN) provides fallback, but its feeder link can become a bottleneck as users migrate. This paper studies post-disaster network resilience through service-priority-aware orchestration of constrained NTN feeder capacity in a post-failure snapshot. A lightweight Cross-Entropy Method (CEM) tunes feeder, offload, and user-ranking parameters offline, not as an online artificial intelligence (AI) controller. A Python system-level simulator sweeps operating regions under abstract TN-NTN link and queueing models. Gains depend on failure severity. At a 180-Mbps feeder under total TN outage, the proposed policy increases the absolute emergency-weighted continuity metric Cemg from 0.399 to 0.666 relative to the matched rule baseline with uniform feeder scaling (paired gain: 0.267; 95% bootstrap CI: [0.264, 0.271]); mild damage can reduce aggregate continuity. Offline CEM feeder-weight tuning further improves severe-failure continuity over fixed service priorities, whereas tuned offload and user ranking provide no detectable incremental gain.

Image-space reward requires decoding and a separate judge; unified visual generators can instead reuse their understanding pathway for decode-free, self-supplied reward. Optimizing against this verifier closes the self-reward loop; shared trainable weights let one update move generation and scoring even with the reward gradient stopped. We call the resulting failure self-verifier capture: self-reward rises while independently measured quality falls. Matched semantic/denoising by shared/frozen cells on Scale-RAE isolate two routes: the policy exploits on-manifold blind spots of a frozen semantic judge, while shared updates additionally rewrite the judge. A frozen judge also becomes stale, and repeated self-refresh yields progressively weaker recoveries. The diagnosis yields a two-part repair: locking the warmed reward read blocks rewriting, while sparse, parameter-disjoint external anchors counter residual exploitation. Unanchored rollouts retain semantic reward across multiple latent noise levels, providing dense latent reward that terminal image-space judging cannot supply, while a drift-triggered controller allocates anchors to high-drift rollouts and propagates their calibration within each group. At matched budgets, the controller improves on fixed-dose anchoring, increasingly so as coverage tightens. Both the unanchored failure and the repair transfer to the higher-performing, architecturally distinct Show-o2. Across both bases, the complete method significantly outperforms a construct-matched every-rollout judge on GenEval and DPG-Bench while externally scoring only 1/16 of rollouts. At 1/32 coverage, it scores higher in all four comparisons and remains jointly non-inferior.

Generative Social Simulation (GSS) leverages Large Language Models (LLMs) to create sophisticated personas that define agents with demographics, emotions, and political opinions. While prior work shows that personas compress outputs into stereotypical caricatures, there is no comprehensive study of how these caricatures evolve into collective artifacts when they interact. Using a Twitter-like GSS platform with personas derived from the American National Elections Study (ANES), we ablate three layers of partisan identity: affective polarization markers, party identification, and vote choice, while retaining demographics, ideological positioning and important issues. We find that collective outcomes depend heavily on how survey responses are translated into personas. Removing a few explicitly partisan statements moves the network from near-complete segregation to cross-party ties approaching those of real Twitter networks. We argue that matching one outcome is not evidence for behavioral validity and point to salience as a possible explanation. Surveys ask for a position on every question regardless of its importance to the respondent, and persona construction turns each answer into a direct behavioral instruction. We conclude with a checklist for more comprehensive evaluation of personas in future works.

In the context of graph reconstruction studies, this paper addresses the problem defined as follows: given a sample of vertex-coloured graphs, each associated with a number, does there exist a connected graph containing exactly the given number of induced subgraphs isomorphic to each graph in the sample, such that every edge of the solution graph is included in at least one of these subgraphs. This paper first shows that this problem is $\Sigma$______-complete. The associated optimisation problem is to determine the minimum number of vertices in a solution graph, if one exists. For instances consisting of two paths, we prove NP-completeness in general and give a polynomial-time dynamic programming algorithm when one path is monochromatic.

This paper investigates the geometric detection of latent dynamic transitions (where the latent space is defined as a finite-dimensional state space reconstructed via a delay-embedding Hankel matrix) in mutually coupled phase-locked loops (PLLs) exhibiting on-off intermittent chaos. We propose a data-driven approach to identify the transition of the effective dimension in latent dynamics as a precursor to physical bursting events. By applying Singular Value Decomposition (SVD) to a Hankel matrix constructed from high-resolution oversampled data, we extract the effective dimension that reflects the dynamical shift of intrinsic variables - from a synchronized two-dimensional manifold to a higher-dimensional chaotic state - prior to the actual burst. This geometric indicator enables a proactive, event-driven intervention using Pyragas-type delayed feedback control, which is activated only when the detected dimension exceeds a critical threshold. Simulation results demonstrate that this geometric precursor effectively compensates for control latencies and substantially reduces accumulated control energy compared to conventional continuous feedback. The findings establish a methodological strategy for stabilizing infinite-dimensional delayed systems by exploiting early warning signals embedded in dimensional transitions of latent dynamics.

We prove that the restarted Arnoldi algorithm for real symmetric matrices with restart length $s=2$ has a cyclic limiting behavior. In the terminology established in (arXiv:2209.14579), this is a proof of the ACI($s$) conjecture for $A=A^T\in \mathbb{R}^{n\times n}$ and $s=2$. As explained in (arXiv:2209.14579), this also proves the Forsythe conjecture (Numerische Mathematik, 11 (1968), pp. 57-76) about the limiting behavior of the conjugate gradient method and positive definite $A=A^T\in\mathbb{R}^{n\times n}$ for $s=2$. The proof we give is an extensively rewritten and simplified version of a proof found using Codex with GPT-5.6 Sol. At the end of this paper, we discuss its origin and evolution, as well as related recent developments around the Forsythe conjecture. Our motivation is to present a proof that is instructive and transparent, so that it can enhance the knowledge of interested human readers.

Improving an agent through successive revisions requires balancing exploration and evaluation under a limited budget. A narrow search can miss useful sequences of edits, while fully evaluating every intermediate version can exhaust the budget before those sequences yield a verified improvement. We introduce Speculative Recursive Self-Improvement (SpecRSI), which drafts executable successor agents before their predecessors are fully evaluated, while requiring complete verification before adoption. Verification-Aware Tree Search (VATS) allocates the budget between further edits and outstanding checks using the predicted verified return of a continuation. Group Verification Optimization (GVO) learns the improver from completed searches while keeping the real evaluator fixed. In the main scientific comparison, terminal gains exceed Janus by 33.3%–40.0%. Additional evaluations show that SpecRSI outperforms Janus in all 50 scientific backbone–task pairs and exceeds MetaRSI-v1 by 1.85%–2.38% in macro-average system score across twenty backbones. Coding-validation success reaches 69.4% versus HGM's 64.3% at a 20.8% lower target-cost index.

The contention is that the current batch of Large Language Models don't *understand* language, but they *do* it. Rather than LLMs forming some kind of “distributed representation” of symbols, the claim is that LLMs are capturing “Cultural Intelligence” and grounding symbols in what the sociologists call practices.

Reinforcement learning (RL) can substantially improve the reasoning capabilities of large language models, but an effective training run requires considerable resources. Reusing the computation acquired through RL across models would extend the value of that training. We propose GrowRL, an RL method that organizes learned updates into compact computational pathways through structural growth. GrowRL freezes the pretrained backbone and adds an independent causal attention branch and a feed-forward network branch at each Transformer layer. Zero-initialized output projections preserve the pretrained policy at initialization, and RL updates only the added pathways. The learned pathways remain explicit after training, allowing transfer to structurally compatible target models and subsequent adaptation with the target backbone frozen. The architecture also supports sequential transfer of additional pathways. By keeping the learned attention and feed-forward computation separate from the pretrained weights, GrowRL provides an explicit structure for preserving and reusing the outcome of an RL run. Experiments across five backbones show competitive mathematical reasoning and out-of-distribution performance with a small number of trainable parameters. Learned pathways provide immediate gains across all three transfer directions, with performance after only 10 target-side updates already close to that after 100 updates in every case. These results support computational pathways as parameter-efficient and reusable units of reinforcement learning adaptation.

Mixture-of-Experts (MoE) models scale model capacity efficiently by activating only a small subset of experts for each input, while still requiring all parameters to be stored in memory, resulting in substantial memory overhead. Expert merging techniques have recently emerged as a promising compression paradigm that reduces this overhead by consolidating multiple experts into a smaller set. However, existing approaches have two key limitations: (i) direct convex interpolation in parameter space is not generally guaranteed to yield meaningful interpolation of expert functions; and (ii) estimating expert similarity or merging coefficients often relies on calibration data, introducing dependence on the calibration distribution. To address these limitations, we propose Spectrum-Aware Latent Expert Merging (SLEM), a calibration-free expert merging framework that occurs in a structured latent representation of expert weights using a VAE. SLEM groups experts within and across layers based on latent similarities and derives merging coefficients from expert weight spectra. It then performs geometry-aware latent interpolation and decodes the resulting representations into merged expert weights. Compared with existing expert merging methods, SLEM achieves comparable or better performance on general knowledge benchmarks and higher average performance on generation benchmarks involving mathematical reasoning and code generation.

Text-to-video (T2V) diffusion models have made remarkable progress in generating high-fidelity visuals, yet their outputs still often exhibit physically implausible phenomena, including discontinuous dynamics and implausible event interactions. Existing representation alignment methods mainly transfer generic video features or short-range motion cues, treating physical plausibility as a single-level representation problem. To address this issue, we propose MiMaAlign, a hierarchical micro-macro alignment framework that transfers complementary dynamics-oriented and event-oriented knowledge into different abstraction depths of a pretrained video diffusion model. Specifically, micro-level continuous dynamics alignment imposes constraints by aligning diffusion features with a continuous-time motion representation learned by a Neural Ordinary Differential Equation (Neural ODE), while macro-level event-commonsense alignment provides representations of object interactions and event semantics through an instruction-prompted vision-language model (VLM). With training-only auxiliary objectives, we align the complementary supervision from the two teachers with shallower and deeper generator features, respectively, enabling the generator to internalize the corresponding physical priors. Empirical evaluations show that, on CogVideoX-2B, MiMaAlign achieves a relative improvement of up to 21.0% in the Physical Commonsense (PC) score on the VideoPhy benchmark over the baseline, and also delivers substantial gains on the highly challenging VideoPhy2 benchmark, while maintaining general video quality on VBench and being supported by blind user comparisons.

Determining crystal structures from powder X-ray diffraction (PXRD) would enable routine characterization of newly synthesized materials, including those absent from reference databases. Generative models offer a promising approach, but existing evaluations remain fragmented and largely assume single-phase samples with known composition. To fill this gap, we introduce PXRD-Bench, a unified benchmark that extends evaluation to unknown per-element atom counts and recovery of both structures in two-phase mixtures. It provides approximately 4.1 million training structures with simulated PXRD, tests generalization to unseen compositions and structural prototypes, and evaluates recovery from experimental PXRD after excluding structurally matching training entries. To provide a strong reference baseline, we also introduce Bragglite, a PXRD-to-structure model developed through a systematic study of PXRD modeling choices. It achieves a 44.2% Top-1 match rate on experimental RRUFF patterns with known composition, 24.6% above the strongest prior baseline. Further analysis yields several key findings: some models fail to benefit from experimental PXRD, while generalization to unseen structural prototypes and recovery from two-phase mixtures remain challenging. Together, our benchmark, systematic studies, and Bragglite clarify the capabilities and limitations of current PXRD-to-structure models.

Deep reinforcement learning has shown strong potential for vision-based autonomous driving, but policies trained in simulation often suffer severe performance degradation when deployed in the real world. Existing sim-to-real methods commonly address this gap through image-transfer-based domain adaptation or policy-agnostic domain randomisation. However, the former requires laborious acquisition of target-domain data and is often limited to the deployment domain, whereas the latter introduces unguided perturbations that may not reflect the underlying sim-to-real discrepancy, leading to suboptimal transfer performance. To overcome these limitations, we propose XFD, an explainable artificial intelligence (XAI)-guided Fourier domain generalisation framework for pure-simulation sim-to-real reinforcement learning. The proposed framework extracts explainability attribution maps from simulated observations, and converts these attributions into policy-conditioned Fourier-domain interventions. By perturbing appearance-sensitive spectral components while preserving task-defining spatial structures, XFD constructs robustness-enhancing visual domains entirely within simulation. Subsequently, the proposed robust policy optimisation guides the policy towards attending to task-critical features while suppressing sensitivity to irrelevant visual information. Extensive experiments in simulation and zero-shot real-world deployment demonstrate that XFD consistently improves traveled distance and reward over standard policy and randomly domain-randomised baseline, especially on unseen maps with cluttered backgrounds. Attribution-based analysis further proves that the proposed policy preserves more consistent task-relevant visual evidence across simulation and real-world observations.

Network simulation plays an important role in HPC/AI system research by enabling exploration of network architectures, communication strategies, and system configurations at scales that may be impractical to study experimentally. Existing approaches to network simulator calibration and validation lack a principled methodology that combines parameter alignment, constrained inference, and independent validation for collective-centric HPC/AI networks. We present CALIPER, a methodology for systematic calibration and validation of HPC/AI network simulators. CALIPER distinguishes directly observable system parameters from auxiliary parameters which are configured via calibration and constrains the latter to feasible domains consistent with the target system. We evaluate CALIPER's hierarchical, scale-aware calibration and its effects on parameter identifiability and generalization using collective communication benchmarks on a reference system with up to 64-nodes at the Jülich Supercomputing Centre. Validation results show that the calibrated parameter configurations identified by CALIPER predict system performance with less than 5% mean errors across scale-up and scale-out network domains. CALIPER demonstrates that hierarchical, constrained calibration can produce simulator configurations whose predictive accuracy persists beyond the observations used for fitting.

Clinical oncology documentation records decisive information in free text, where it remains readable but not retrievable at scale, leaving cohort identification dependent on incomplete diagnostic codes or manual review. The current standard, coding individual terms, captures what a document mentions, not what it asserts: meaning lies in how those terms relate to one another. We present a method for deriving queryable knowledge graphs from clinical narratives, conforming to the Extended Grammar of SNOMED CT across multiple ontologies. The method specifies how entities should be annotated and related. It defines five relation classes, each bound to a construction rule that assembles the entities and relations into post-coordinated expressions. A deterministic implementation, UNIGRAPHE, applies those rules to construct knowledge graphs from annotated text. Applied to expert-annotated real-world oncology documentation, UNIGRAPHE reconstructed 5,299 annotations into 2,196 post-coordinated expressions and shows that relation structure is the missing piece to make clinical narrative formally queryable.

Fix a finite alphabet $\Sigma$ having at least two letters, and natural numbers $n,k\geq 1$. We denote by $D_n(\mathbf{\Pi}^0_k)$ the class of differences of $n$ subsets of the Cantor space $\Sigma^\omega$ in $\mathbf{\Pi}^0_k$, and by $\checked{D}_n(\mathbf{\Pi}^0_k)$ the class of complements of sets in $D_n(\mathbf{\Pi}^0_k)$. We prove, for any natural numbers $n,k\geq 1$, the existence of a finite alphabet $\Sigma$ having at least two letters and a finitary language $U_{n,k}$ over $\Sigma$ accepted by a one-counter automaton such that the $\omega$-power $$U_{n,k}^\infty:=\{ w_0w_1\cdots\in\Sigma^\omega\mid\forall i\in\omega ~~w_i\in U_{n,k}\}$$ is $D_{2n-1}(\mathbf{\Pi}^0_k)$-complete. We also prove a conditional result for the dual classes $\checked{D}\_{2n}(\mathbf{\Pi}^0\_k)$: under the assumption that there exists a finitary language $V_{n,3}$ accepted by a one-counter automaton, such that its $\omega$-power $V_{n,3}^\infty$ is $\checked{D}_{2n}(\mathbf{\Pi}^0_3)$-complete, there would also exist, for any integer $k>3$, a finitary language $V_{n,k}$ accepted by a one-counter automaton, such that its $\omega$-power $V_{n,k}^\infty$ is $\checked{D}_{2n}(\mathbf{\Pi}^0_k)$-complete. Finally we add another conditionnal result about Wadge degrees of rational $\mathbf{\Delta}^0_2$ omega powers, which completes the recent results of [FGL26].

Video anomaly detection (VAD) identifies deviations from normal patterns, but such deviations do not necessarily correspond to safety-critical events. We focus on human-centered emergencies and violence and investigate whether motion-based anomaly discrimination reflects task-relevant motion information or observation-quality differences arising from person detection, pose estimation, and tracking. We introduce an observation-quality-aware evaluation framework that combines quality diagnostics, quality balancing, motion orthogonalization, and robustness analyses. On human-centered categories of UCF-Crime, quality-only features match or exceed the discrimination of multi-cue motion in the original cohort, revealing a strong observation-quality shortcut. After cardinality matching brings all 11 measured quality variables below a prespecified marginal balance threshold, multi-cue motion consistently outperforms quality-only features across Logistic Regression, Random Forest, and RBF-SVM, yielding ROC-AUC gains of 0.046-0.141. Because nonlinear quality-only prediction remains above chance after matching, we interpret this result as additional discrimination beyond measured marginal quality rather than as evidence of quality-free or causal motion understanding. These findings show that high VAD performance should not be interpreted as evidence of abnormal-motion modeling without auditing the reliability of the person-observation pipeline. Our study provides a systematic protocol for evaluating observation-quality shortcuts and for assessing whether motion-based discrimination survives explicit control of detection, pose-estimation, and tracking reliability.

White-box neuron pruning has shown that refusal in large language models (LLMs) can depend on a small, removable set of feed-forward (FFN) channels. Existing attacks either need carefully paired “twin” prompts or score each neuron in isolation. Pruning also has an unavoidable capability cost, so over-removal compounds the loss. TwinBreak removes $1\%$–$5\%$ of parameters and NeuroStrike removes $0.3\%$–$0.7\%$ of neurons in targeted layers. We model safety as a small circuit whose neurons contribute jointly rather than independently. We propose CircuitBreak, a neuron-pruning attack that greedily searches for a sparse set of FFN neurons whose joint ablation flips a refusal readout. CircuitBreak re-scores every candidate after the neurons already removed and needs no twin prompts. On ten instruction-tuned models, the greedy search converges after removing $29$–$150$ neurons ($0.0066\%$–$0.0388\%$ of the MLP). The pruned models raise the attack success rate from $2.24\%$/$7.35\%$ to $87.54\%$/$87.54\%$ on Llama-3.2-3B-Instruct/Llama-3.1-8B-Instruct and from $8.95\%$/$5.43\%$ to $88.50\%$/$88.18\%$ on Qwen2.5-7B-Instruct/Qwen2.5-14B-Instruct, as judged by Llama Guard 3 8B on StrongREJECT.

Procedural guidance can show users how to act without explaining what an action enables later. We present Foresee, an interface combining next-step guidance with consequence previews, method comparisons, and cross-step access. A hybrid pipeline combines structured rendering, image editing, and video generation. A literature-informed framework and six-participant formative study shaped its design. A 24-participant study compared AR guidance, AR with generated video, and Foresee across digital painting, gameplay, and cocktail making. Exploratory painting comparisons showed higher correct completion without a recorded error than AR and higher post-task quiz accuracy than generated-video guidance under task-wise correction. Interviews revealed that users could overlook valued explanations, match visible outcomes without understanding later possibilities, or wait for concrete demonstrations when symbolic cues were unclear. These findings inform generative previews that explain what actions enable or constrain while supporting ongoing execution.

We present PatMT2 (Patent Machine Translation, v2), a two-stage LoRA pipeline that adapts the Hy-MT2-7B dedicated translation model to four-language (Chinese/English/Japanese/Korean) patent translation. On six translation directions, the Stage-2 checkpoint reaches a macro-average BLEU of 43.39 (BF16) to 43.57 (AWQ INT4), improving over the Hy-MT2-7B base by up to +8.8 BLEU and also exceeding Qwen3-32B and several external systems on the same average. We then measure the cost of adaptation: probing 38 languages absent from the fine-tuning data, the model still produces correct, fluent output on 34, while four low-resource languages collapse into distinct, reproducible failure modes. We treat these as diagnostic cases rather than retention rates, and report their mechanisms without attributing them to a single cause. Paired bootstrap tests confirm the translation gains over all baselines (p < 0.001).

Vision-language models answer questions about images they were never shown, a failure called mirage. Such an answer can be correct without being justified by the evidence, and the difference matters wherever a model's output is checked by someone who did not see its input. We locate the failure in the training objective: cross-entropy separates a model that reads the image from one that ignores it by only the conditional mutual information between answer and image, and no training example rewards declining. We formalise support completion, TIDE, which also shows each training question with a blank image, a same-domain wrong image and an out-of-domain image, with targets that decline, and prove that under TIDE every image-blind predictor is suboptimal by at least 1.04 nats per example, that the optimum declines exactly on the counterfactual kinds while keeping its behaviour on valid images, and that for any finite model the error of its answer-or-decline decision, declining on valid input plus answering on invalid input, is bounded below by how well its representation separates valid from invalid evidence. On Qwen2.5-VL-7B under a frozen held-out protocol with two judges, the three-condition predecessor of TIDE reduced the invalid-evidence mirage rate from 0.759 to 0.097 and the blank-image rate from 0.700 to 0.003 at a false-abstention cost of 0.126; label-free grounding raised the rate to 0.859, a likelihood-advantage contrast moved it only to 0.718 because its empty-image term carries no gradient, and grounding applied after support completion re-opened the mirage. The four-condition protocol on three base models, with seeds and component ablations, is in progress and not reported here.