Papers

Harness design has transformed the development of LLM-based agents by integrating memory, tool use, and runtime control. However, this design also introduces security and privacy risks because malicious instructions from external sources may be written into persistent memory and persist across sessions. To study this risk, we propose PMPA, a Persistent Memory Poisoning Attack against harness-based agents. PMPA embeds malicious instructions into benign external sources and induces the victim agent to write them into persistent memory without directly accessing to the agent framework. Once stored, the poisoned memory can be retrieved in later sessions, triggering additional malicious actions and causing privacy leakage. We evaluate PMPA on OpenClaw and Claude Code across different backbone LLMs, input modalities, and trigger scenarios. Across all settings, PMPA achieves average Injection Success Rate (ISR) and Cross-session Attack Success Rate (C-ASR) of 73.7%/ 55.5% on OpenClaw and 66.9%/ 81.7% on Claude Code, while preserving benign task performance on both systems. We further evaluate a targeted prompt-level defense and find that it can reduce memory injection in many settings, but provides limited protection once the persistent memory has been poisoned.

We study a finite-state partially observable restless multi-armed bandit (PO-RMAB) motivated by resource-constrained wildlife monitoring. The underlying condition of each location evolves independently, while only a limited number of locations can be actively monitored at each decision epoch. Activation reveals the current state, whereas passive operation provides no observation, yielding a collapsing-bandit belief dynamics. Our main contribution is a structural characterisation of optimal policies for the multidimensional belief-state problem. We establish sufficient conditions under which an activation advantage is monotone with respect to the belief state, and hence the optimal policy has the threshold type structure over the $(M-1)$-dimensional belief simplex. The key result is obtained by bounding the variation of the value function through a model-dependent Lipschitz constant. We further specialise the result to one-step birth-death dynamics and derive explicit bounds.

This paper introduces adaptive moving mesh methods for the numerical simulation of axisymmetric mean curvature flow, addressing both isotropic and anisotropic cases. The methods are developed within the framework of the mesh equidistribution principle, where a carefully designed tangential velocity is employed to dynamically redistribute mesh points during the evolution. To accurately capture the key geometric features of the evolving interfaces, we select monitor functions based on the curvature $\kappa$, its arc-length derivative $\kappa_s$, and the squared curvature $\kappa^2$. These monitor functions can be flexibly tailored to suit different problem settings and play a vital role in determining the resulting mesh quality and numerical accuracy. Spatial discretization is performed using central finite differences, while temporal integration is handled with first- and second-order time-stepping schemes, including the BDFk ($k=1,2$) and Crank-Nicolson methods. Additionally, a Lagrange multiplier approach is incorporated into the adaptive system to enforce the underlying geometric constraint, resulting in energy-stable numerical schemes. Numerical experiments confirm the convergence and energy stability of the proposed methods. More importantly, the results clearly show that the proposed methods offer significant advantages in complex geometric evolutions: by utilizing appropriately designed monitor functions, the adaptive methods achieve dynamic redistribution of mesh points, efficiently capturing localized geometric features, significantly improving numerical accuracy, and effectively preventing mesh degeneration, particularly in anisotropic cases.

Studio recording techniques vary considerably, ranging from live recordings in a shared acoustic space, to isolated booth recording and overdubbing, where musicians record their parts separately, while listening to previously recorded material. These approaches differ in terms of physical co-presence, visual contact, sound leakage, and acoustic isolation, factors that may influence musical coordination. This study investigates whether the recording style affects timing precision; specifically, whether musicians performing together in the same space achieve tighter temporal coordination. To address these questions, two datasets with a total of 391 multi-track recordings were analyzed to measure the timing relationships between musicians. In addition, an automated sound-leakage detection method was employed to infer the recording conditions of each song: pairwise Mel-spectrogram similarities between the tracks were computed, and summary statistics from the resulting similarity matrices were used to identify recordings with shared acoustic content. The results suggest that recordings showing evidence of shared room acoustics and sound leakage tend to exhibit lower timing variability than highly isolated recordings, whereas overdubbing is associated with increased timing variability.

An acquired representation can enlarge a system's cognitive repertoire without transferring the capacities exercised in producing that representation. This paper develops a framework for specifying that enlargement and its limits. Its central contribution is a five-part attribution table distinguishing effective tracking, application of acquired structures, acquisition from explicit specifications, acquisition from identifying observations, and retention and reuse. Each entry identifies a positive capacity commitment and a further claim requiring additional support. The argument deliberately grants meaningful content, causal efficacy, and productive inference, so that its conclusion does not depend on treating representations as inert encodings. Map and category examples show why even complete application competence leaves acquisition capacity undetermined, and why acquiring a criterion from its description differs from finding it in examples. Short formal proofs appear in an appendix. The framework is applied to Andrew Ng's world-model interpretation of Othello-GPT and to the specific indicators discussed in contemporary accounts of machine concepts. It preserves demonstrated recognition, classification, inference, and qualified acquisition while specifying what remains unestablished about criterion discovery and accumulation. The result concerns the scope of cognitive attribution rather than the constitutive conditions of concept possession: cognitive achievements deserve credit for the capacities they establish, without silently importing a broader repertoire through the labels attached to them.

Spatio-temporal traffic data are central to intelligent transportation systems, yet their heterogeneity poses significant challenges for large-scale modeling. Existing pre-trained models often rely on a homogeneous modeling paradigm to handle highly heterogeneous traffic data. This fundamental mismatch not only limits model generalization but also leads to computationally expensive and parameter-inefficient designs. To this end, we propose FlexST, a novel pre-training framework that introduces modularity and adaptivity for traffic modeling. Specifically, we first propose a multi-resolution spatio-temporal diffusion module that captures both short-term fluctuations and long-range trends, effectively reconciling inputs with divergent temporal and spatial resolutions. After that, we construct a domain-adaptive mixture-of-experts that dynamically routes data to specialized sub-networks, enabling selective knowledge transfer while preventing negative interference across diverse domains. Moreover, we devise a unified periodic encoding strategy that injects resolution- and domain-aware inductive biases to harmonize periodic inconsistencies across datasets. Extensive experiments on 23 real-world traffic datasets demonstrate that FlexST significantly outperforms state-of-the-art baselines in zero- and few-shot settings, showcasing superior generalization, adaptability and efficiency. This work offers a new direction for building general-purpose pre-trained models capable of handling the complexity and variability of urban traffic systems.

Temporal Path Cover (TPC) and Temporally Disjoint Path Cover (TDPC) ask for minimum-cardinality covers of a temporal digraph by temporal paths, with TDPC additionally requiring pairwise temporal disjointness. Cioni et al. proved both problems NP-hard on temporal DAGs whose underlying undirected graph has vertex-cover number three and left the vertex-cover-two case open. We resolve this question affirmatively for both problems. Using polynomial-time reductions from Distinct Numerical Matching with Target Sums, we show that TPC and TDPC are NP-complete already when the underlying undirected graph has vertex-cover number exactly two. For TPC, this yields an exact vertex-cover threshold: polynomial-time solvability at vertex-cover number at most one and NP-completeness at two. For TDPC, we further study the boundary case of temporal oriented stars. A reuse-normalization argument reduces optimum solutions to leaf-unique two-edge merges. If either the incoming or outgoing side is single-label, the optimum remains polynomial-time computable even when the opposite side is multi-label. We also introduce a span parameter for feasible merge realizations and prove that any feasible family of k merges forces a realization of span at least 2k-2. Hence TDPC on temporal oriented stars is in XP parameterized by maximum span, and is polynomial-time solvable for every fixed span bound. The unrestricted multi-label star case remains open.

Personal robots should adapt their co-speech gesture style to a new user without requiring model retraining. We present a training-free personalization pipeline that combines a frozen audio-conditioned diffusion prior with a gesture style encoder and lightweight conditioning adapters. The encoder is first trained to discriminate speaker identities and then jointly refined with the adapters using the diffusion objective, enabling a reusable style embedding to be extracted from approximately 10 seconds of enrollment motion through a single forward pass. To support this setting, we also release a Quest~3 capture application and a dataset of spontaneous co-speech motion from ten participants. We evaluate the system on held-out speakers using Style Recognition Accuracy (SRA) and Fr'echet Gesture Distance (FGD) to measure personalization and motion quality. Our approach improves SRA from 27.6% for the frozen prior to 69.5% while preserving motion quality (FGD 34.2 versus 34.8), and replacing the enrollment embedding with another person's reduces SRA to 11.4%. The generated gestures are retargeted to a physical NAO robot, and this improvement also transfers to the robot deployment setting, where speech is synthesized using five TTS voices, retaining 67.3% SRA.

We introduce cvxgenrust, an open-source tool for generating custom Rust code that solves families of parameterized convex optimization problems modeled in CVXPY. cvxgenrust canonicalizes a problem family, extracts affine maps to Clarabel cone-program data, and generates a specialized Rust crate that updates parameters and calls Clarabel natively at runtime. The generated solver can also be exposed to Python and registered as a custom CVXPY solver. Our code generator supports a wide range of convex optimization problems up to semidefinite programs and exponential-cone problems. Numerical experiments show reduced runtime relative to direct CVXPY solves and performance comparable to CVXPYgen on shared problem classes.

Multimodal clinical AI typically assumes that the waveform, report, metadata, and downstream predictions attached to a record belong to the same patient. In practice, linkage failures can silently assemble individually plausible but cross-patient components, creating a safety problem that standard predictive models are not designed to detect. We study this problem as multimodal record integrity triage: given an assembled record, should its modalities be trusted to belong together? We introduce SHIFT-M3, a lightweight text-based pre-fusion screen that measures alignment-based consistency between two separately produced ECG text views: an LLM-generated interpretation and a clinical report summary. On 784,680 MEETI ECG records, SHIFT-M3 achieves 97.6% TPR@5% FPR for full text-view swaps (AUROC 0.996), 90.3% for partial swaps (AUROC 0.974), and 97.7% for label-matched hard negatives (AUROC 0.996) with only 573,569 parameters. Compared with same-dataset lexical baselines, the gains are largest on partial swaps and hard negatives, suggesting that the model is learning more than surface overlap. We also introduce the CMST (Conflict-type Multimodal Stress Test) evaluation taxonomy, a three-seed stability study, a loss ablation, a temporal-tolerance sweep, and a shared-token masking control. The main remaining failure mode is longitudinal ambiguity: at the default operating point, same-patient cross-visit pairs still produce 87.0% Type-II false positives.

Mobile GUI agents increasingly rely on Vision-Language Models (VLMs) to automate smartphone tasks by interpreting screenshot streams. However, this design introduces serious and underexplored privacy risks, including direct leakage of sensitive on-screen information and unintended user profiling. The absence of standardized benchmarks makes it difficult to quantify these risks in realistic mobile agent workflows. To address this gap, we propose PriMobiBench, the first benchmark for systematically evaluating privacy leakage and visual profiling in screenshot-driven mobile agents. It provides a unified pipeline for data generation, agent trajectory construction, and multi-model evaluation. We also introduce MobiLeak, a dataset of execution traces from 16 apps, covering 25 privacy attributes with 2,960 embedded privacy instances. Our results reveal substantial risks: (1) VLMs can directly extract sensitive information with up to 82.5% success rate; (2) beyond explicit leakage, they can infer user profiles from aggregated visual evidence with approximately 70% success. We further propose a mitigation that masks privacy-sensitive but task-irrelevant UI elements before cloud processing, reducing profiling success by up to 58% with only approximately 8% performance loss. Overall, our work provides the first systematic benchmark for visual privacy risks in mobile GUI agents, demonstrates that both leakage and profiling are feasible at a highly concerning level, and offers a practical direction for mitigation.

As 6G networks transition from theoretical frameworks into operational realities, artificial intelligence (AI) evolves from an add-on optimization tool into a distributed and interconnected structural layer. Unlike previous network generations that mostly relied on centralized cloud analytics platforms, AI-native 6G networks operate across a dynamic, multi-domain edge-cloud continuum where data originates from heterogeneous sources including user devices, radio access networks, sensing infrastructures, and vertical applications. Centralizing this massive volume of data creates severe communication overhead, unacceptable latency bottlenecks, single points of failure, and complex cross-domain governance challenges. Consequently, decentralization becomes a fundamental architectural requirement for future 6G network intelligence and zero-touch operations. Security serves as the primary enabler of this decentralized paradigm. Critical security functions, such as real-time threat detection, physical-layer attack mitigation, slice protection, and intrusion detection, require immediate access to local context and telemetry before operational data loses its value. However, moving intelligence to the edge via collaborative paradigms like federated learning (FL) and decentralized FL (DFL) introduces complex trade-offs. System security cannot be addressed in isolation; it is deeply intertwined with equally important aspects like trustworthiness, explainability, and energy sustainability. Taking these aspects into account, this paper develops a unified perspective on decentralized intelligence for 6G, arguing that decentralization, trustworthiness, explainability, and sustainability must be designed jointly rather than treated as independent requirements

The language $\mathrm{MIX}$ consists of all words over a three-letter alphabet that have an equal number of occurrences of each letter. It is also the word problem of $\mathbb{Z}^2$ with respect to a suitable choice of generators. The Kanazawa–Salvati conjecture states that $\mathrm{MIX}$ is not a well-nested multiple context-free language. Every well-nested multiple context-free language is an indexed language. We reduce the conjecture to an explicit combinatorial problem about tuples of words, which is easier to state than the original formulation in terms of arbitrary well-nested multiple context-free grammars. More generally, for every surjective monoid homomorphism $\psi\colon \Sigma^* \to \mathbb{Z}^d$, we define a family of well-nested multiple context-free grammars $G_\psi[r]$ for $r \geq 1$, each of which generates a sublanguage of $\psi^{-1}(\mathbf{0})$. We prove that every well-nested multiple context-free sublanguage of $\psi^{-1}(\mathbf{0})$ is contained in $L(G_\psi[r])$ for some $r \geq 1$. Using this family, we prove that the four-letter analogue $\mathrm{MIX}_4$, which is a word problem of $\mathbb{Z}^3$, is not a well-nested multiple context-free language. The proof reduces this claim to a result of Bishop–Elder–Evetts–Gallot–Levine stating that the two-letter analogue $\mathrm{MIX}_2$ is not generated by any non-branching multiple context-free grammar. The Kanazawa–Salvati conjecture itself remains open.

Automatic sentence function identification is important for many downstream natural language processing (NLP) applications such as dialogue systems, text-to-speech synthesis, and machine translation. However, benchmark resources for Bangla sentence function classification remain limited. To mitigate this gap, this paper introduces a corpus of 10,000 Bangla sentences, manually annotated into four functional categories, namely declarative, interrogative, imperative, and exclamatory. The corpus is nearly balanced across the four classes, with high annotation reliability reflected by a FleissḰappa of 0.82. Furthermore, we evaluate multiple feature representations, including Bag-of-Words (BoW), TF-IDF, and Word2Vec, with several classical machine learning classifiers. In addition, two heterogeneous ensemble models, namely Single-Level Ensemble (SLE) and Double-Level Ensemble (DLE), are utilized to improve classification performance. Experimental results show that TF-IDF consistently outperforms Word2Vec, likely due to its ability to emphasize discriminative lexical cues associated with sentence functions, particularly given the relatively small corpus used to train Word2Vec. The DLE model with TF-IDF features achieves the best performance with accuracy and macro-F1 of 0.95, demonstrating the effectiveness of sparse lexical representations and heterogeneous ensemble learning for this task. Further cross-validation confirms the robustness of the approach, while LIME-based interpretability provides insights into model predictions. The developed corpus and model benchmarking establish strong baselines for Bangla sentence function classification.

Earth-observation (EO) foundation models have become exceptionally effective at learning se mantic, high-dimensional geospatial embeddings, while modern weather and climate models have demonstrated that Earth-specific geometry, spherical operators, meshes, and hybrid physical solvers can materially improve prediction. Yet these two advances are not equivalent. A conventional latent embedding has no inherent physical transformation law, whereas scalar fields, tangent polar-vector fields, axial/pseudovector quantities, covectors, and higher-order tensors transform differently under rotations, reflections, and changes of local coordinate frame. This proposal asks whether a general purpose Earth foundation model should preserve those distinctions explicitly, or whether standard embeddings plus augmentation already learn everything that matters. The central contribution is therefore not a more complicated architecture by assumption, but a staged falsification program. A compute-conscious ERA5 dry run first compares conventional, augmentation-matched, typed equivariant, and Hodge/Helmholtz variants under spatial, temporal, orientation, and low-data shifts. Only if explicit geometric typing yields reproducible improvements does the program advance toward a multimodal Earth foundation model in which semantic embeddings coexist with physically typed fields. The proposed gap is narrower and more defensible than claiming that current models ignore geometry entirely: several systems already respect spherical domain geometry, and emerging work explicitly learns scalar/vector fields on spheres. The unresolved question is whether foundation-scale, multimodal, parity-aware field typing produces practical gains beyond those existing approaches.

Social networks shape how individuals make decisions and how opportunities are distributed. However, the mechanisms that generate these networks often reflect pre-existing inequalities, and technologies that rely on network-derived signals risk further amplifying such disparities. Algorithmic fairness research largely treats networks as a fixed background, grounding analysis almost exclusively in distributive justice and overlooking how network structures systematically bias decision-making. In this Perspective, we identify ten network effects and trace how they create structural biases in the relationship between what we intend to measure and what we observe. Using academic hiring as an example, we show that network biases are not inherently harmful or beneficial. Determining their legitimacy requires examining the entire decision-making process through the lenses of both distributive and procedural justice while engaging all affected stakeholders. We therefore call for a holistic, networked approach to fairness that moves beyond static group categories and recognizes the dynamic, relational, and structural nature of inequality.

Group-relative policy optimization (GRPO and descendants) can discard no-contrast rollout groups through dynamic sampling, while practical implementations expose a configurable filter metric. We identify and quantify a metric-predicate mismatch under composite shaped rewards. When filtering follows the shaped training score rather than the task outcome, all-fail groups retain nonzero within-group spread and pass the predicate; standard-deviation normalization then promotes shaping differences among failures to full-size phantom advantages. In a controlled GSM8K comparison (Qwen2.5-1.5B, LoRA), no filtering and shaped-score filtering end at EM 0.080 +/- 0.112 and 0.040 +/- 0.008, whereas binary-outcome filtering holds 0.754 +/- 0.005 across four runs per arm (three default-seed reruns and one seed-123 run; mean +/- sample SD). On verl's native recipe/dapo trainer, holding model, data, reward and trainer fixed and changing only the metric, the score arm requires no batch refill in any of 40 observed steps and ends at EM 0.160; the accuracy arm refills in 29/40 steps and ends at 0.763. Both use the same custom shaped-reward hook and unmodified trainer/filter code. Prior work established shaping-induced amplification and all-fail filtering; our contribution isolates the metric-predicate semantic mismatch and directly instruments native deletion/refill telemetry. Across tested positive coefficients lambda in {0.1, 0.3, 0.5}, unsafe arms collapse; exploratory one-run cells reproduce the failure at 1.5B/7B on MATH and under GSPO, while disabling standard-deviation normalization avoids the observed collapse. Filtering under a composite reward should use a task-outcome signal whose semantics are independent of shaping.

Agriculture accounts for roughly 70% of global freshwater withdrawals, yet irrigation is still commonly scheduled reactively, with no forecast of where soil moisture is heading and no statement of confidence in that forecast. Data-driven models are accurate but opaque and point-valued; water-balance models are transparent but carry large structural error. Neither alone supports a defensible irrigation decision under uncertainty. This study coupled the two and carried uncertainty through to the decision: a four-parameter water-balance core, calibrated on training data only, was corrected by a Random Forest that learned nothing but the physical residual, conformal prediction attached 90%-nominal intervals, and a risk-aware rule converted the interval lower bound into an irrigation trigger. It was evaluated on three years of hourly in-situ measurements from a rainfed Mediterranean cropland station under a strict chronological split, scored against persistence, from one hour to one week. At the 24 h horizon the hybrid reached RMSE 0.00925 m^3 m^-3 and +9.4% skill, roughly double the best of nine baselines, of which only the Random Forest beat persistence. Skill did not grow with lead time: it peaked at +27.4% at three hours and fell to +1.2% at one week. Conformalised quantile regression was better calibrated and 11% sharper than constant-width conformal prediction. The risk-aware rule raised management-threshold crossings detected in advance from 0.905 to 1.000, at a precision cost of 0.975 to 0.950 and 3.7% more notional water, and beyond 72 h the point forecast fell below the no-forecast rule while the interval-based rule did not. Uncertainty quantification therefore governs the lead time over which forecast-driven irrigation advice remains trustworthy, and here transparency in the physical layer cost no measurable accuracy.

Querying clinical trial registries remains a manual and error-prone process, requiring researchers to navigate large volumes of semi-structured data without support for natural language interaction or cross-source synthesis. To address this, we introduce ClinAgent, a conversational system based on agentic Retrieval-Augmented Generation (RAG) that enables clinicians and researchers to query clinical trial information in plain language and receive grounded, up-to-date responses across multi-turn interactions. The system centers on a Large Language Model (LLM) agent following the ReAct paradigm, which iteratively reasons over queries, selects among a set of integrated tools, and refines its actions based on intermediate outputs. These tools include a ClinicalTrials.gov search interface, a PubMed module, and a Python-based analyzer operating on a locally cached structured dataset of clinical trials. We evaluate the system using a three-phase framework assessing operational effectiveness, planning quality, tool-use efficiency, and expert qualitative judgments, comparing three LLM backends: Gemini 3.0 Flash and two variants of DeepSeek V3.2 (thinking and non-thinking). Results reveal complementary strengths, with DeepSeek (thinking mode) excelling in planning quality, while Gemini achieves the highest overall performance and strongest expert ratings. Overall, our findings highlight the potential of agentic AI systems to improve the accessibility and synthesis of clinical trial information, supporting more efficient and user-centered biomedical research workflows.

Deploying reconstructed 3D environments on standalone VR headsets are constrained by limited compute and memory, and conventional level-of-detail policies optimize for visibility without accounting for the user's explicit inspection intent. We present GaussAnything, a native OpenXR system for intent-conditioned reallocation and progressive publication of evolving semantic Gaussian+SDF scenes. GaussAnything resolves class- or instance-level queries to persistent 3D objects and reallocates a fixed Gaussian resident budget toward the selected object while retaining global context, applying incremental, stable-identity updates coordinated with the TSDF-derived mesh through a source-epoch mechanism. Across eight scenes, an object query concentrates 88-90% of the fixed client budget onto the queried object without enlarging it, on-device rendering reproduces the host render to within a small margin (up to 36.7 dB), and the standalone client renders each stereo frame at a steady-state GPU cost of roughly 10 ms within the frame budget of standard standalone panels.

Multimodal emotion recognition in conversation (ERC) requires adapting to the instance-dependent reliability of different evidence sources. Lexical content may be decisive, vocal expression may provide complementary cues, or accurate recognition may require cross-modal interaction; fixed fusion does not explicitly account for this variation. We propose ReH-FUSE, a reliability-aware framework with dialogue-aware text, audio, and cross-modal experts. Its decision-level router first models the relative preference between text and audio and then balances the resulting unimodal mixture against the cross-modal expert. This factorization separates unimodal competition from cross-modal selection. Across three independent runs on IEMOCAP, ReH-FUSE achieves 74.34% weighted F1 and 73.11% macro F1; on MELD, it achieves 68.03% weighted F1. Controlled ablations show that learned routing outperforms uniform expert averaging and benefits from cross-modal interaction.

Enterprise customer support systems must answer customer questions correctly, retrieve the right policy information, use customer context, and pass difficult cases to human agents when needed. This paper presents ShopEase, a Generative AI-based multi-agent framework for enterprise customer support. The system combines six components: Intent, CRM, Memory, Hybrid RAG, Escalation, and Supervisor, and uses LLaMA 3.2 running locally through Ollama for response generation. The retrieval module combines FAISS (dense retrieval) and BM25 (sparse retrieval), and six configurations are evaluated: BM25-only, FAISS-only, Fair RRF, Weighted RRF, RRF with Cross-Encoder, and Top-10 Hybrid with Cross-Encoder. Instead of using a fixed mapping between intent and policy, the policy category is decided directly from the retrieved documents. The system was evaluated on 2632 held-out customer queries across six categories: Refund, Return, Shipping, Cancellation, Damaged Product, and Unknown. FAISS-only achieved the highest accuracy of 85.37% (2247 correct predictions), closely followed by Weighted RRF at 85.07%. BM25-only achieved only 55.74% accuracy. Adding cross-encoder reranking did not improve results: RRF with Cross-Encoder reached 83.24%, and Top-10 Hybrid with Cross-Encoder reached 81.88%, while also increasing response latency. Category-level analysis shows strong performance on Shipping, Cancellation, and Return, while Unknown queries remain the main source of errors. Statistical testing using McNemar's test shows no significant difference between FAISS-only and Weighted RRF, though both perform significantly better than Fair RRF and the cross-encoder configurations. Overall, dense retrieval gives the best accuracy on this dataset, and additional reranking adds processing time without improving classification performance.

Facial emotion recognition (FER) in real-world environments remains challenging due to unconstrained imaging conditions, including multiple faces, occlusions, pose variations, and complex lighting. Most existing studies focus on individual facial emotion classification and do not address the analysis of collective emotional states at the event level. This paper proposes an end-to-end pipeline for event-level emotion recognition from photographs. The approach detects faces in each image, classifies facial expressions using a deep convolutional neural network, and aggregates face-level emotion probabilities to estimate the overall emotional distribution of a public event. A comparative evaluation of several CNN architectures on the FER- 2013 and RAF-DB datasets demonstrates that transfer learning with EfficientNet-B2 trained on RAF-DB is more suitable for real-world RGB data. The proposed method is evaluated on a real-world event dataset containing 1658 images. Experimental results show stable emotion distributions across event subsets, confirming the effectiveness of event-level aggregation for emotion analysis in the wild.

Post-training vision-language-action (VLA) models for specific robots and tasks requires in-domain demonstrations, yet collecting diverse robot data is costly. Egocentric human demonstrations provide a scalable alternative, but directly mixing human and robot data can introduce cross-embodiment discrepancies and degrade policy performance. To address this challenge, we introduce ReWeight, a framework that incorporates human data into VLA post-training through demonstration-level retrieval and sample-level weighting. ReWeight learns a cross-embodiment visuomotor representation that combines visual observations with future actions to measure behavioral similarity between human and robot demonstrations. Based on optimal transport, it retrieves human demonstrations relevant to the target robot data and assigns larger weights to samples with smaller cross-embodiment discrepancies. We evaluate ReWeight using $\pi_{0.5}$ across eight simulation tasks and four real-world tasks under both clean and randomized settings. In simulation, ReWeight improves the average success rate of post-trained $\pi_{0.5}$ from 39% with only robot data and 44% with randomly mixed human-robot data to 57%. In the physical experimental setting, it achieves an average success rate of 68.8%, outperforming the baselines by 28.8% and 13.8%, respectively. Overall, ReWeight provides an effective paradigm for transforming abundant egocentric human experience into transferable supervision for robot learning. (Project webpage: https://reweight-vla.github.io/)

Multimodal Large Language Models (MLLMs) often struggle with complex mathematical visual reasoning primarily due to a lack of fine-grained perception, causing initial visual hallucinations to directly trigger cascading reasoning failures. In traditional end-to-end reinforcement learning (RL), sparse rewards fail to decouple perceptual hallucinations from logical missteps, hindering targeted perception optimization. Alternatively, fine-tuning with perception-enhanced CoT data incurs high costs and hallucinations. In this paper, we address these challenges by proposing UniCAR-RL, an annotation-free RL framework. By explicitly decoupling the optimization of perception and reasoning during the training process, it achieves isolation and optimization of both capabilities. Specifically, UniCAR-RL consists of three synergistic branches: 1) a Caption-RL branch that optimizes perception capabilities through verifier-guided reasoning validation; 2) a Reasoning-RL branch that performs logical reasoning based on a gold image description to halt cascading errors; 3) a QA-RL branch that retains native end-to-end alignment to ensure robust question-answering performance. Experiments show that UniCAR-RL substantially improves MLLMs' mathematical and visual reasoning using only raw short-answer data. Furthermore, it demonstrates strong generalization across diverse architectures and scales.

In this paper, we extend FLORES+ with Portuguese-source evaluation sets for three Mozambican Bantu varieties: Xichangana, Mozambican Nyanja, and Sena. We compare Xichangana with the existing Tsonga reference and Mozambican Nyanja with Chichewa, and evaluate NLLB-200, Google Translate, GPT, and a variant-aware NLLB model. Holding system output fixed reveals substantial reference sensitivity. On devtest, changing only the reference from Tsonga to Xichangana reduces spBLEU by 13.10 points for NLLB-200 and 15.30 for Google. On matched Nyanja subsets, replacing Chichewa with Mozambican Nyanja produces smaller but consistent reductions of 3.03 and 6.10 spBLEU, respectively. Variant-aware fine-tuning reverses this pattern on the intended targets: relative to NLLB-200, it improves Xichangana by 7.04 spBLEU and Mozambican Nyanja by 5.33 on devtest, while losing performance on the sibling references. GPT is competitive on Tsonga and Chichewa but substantially weaker on the Mozambican varieties. For Sena, the finetuned model reaches 12.64 spBLEU and 36.21 chrF++ on devtest. These findings motivate variety-aware language identifiers, references, and reporting for cross-border languages or language dialects/variants. The data is publicly available on Hugging Face at https://huggingface.co/datasets/MOZNLP/FLORES_MOZ

Delayed Tensor Parallelism (DTP) removes the blocking all-reduce of tensor-parallel Transformer inference. Every device adds its own partial output to its residual stream (and broadcasts it) immediately, but only gathers (receives) the other devices' partials $\delta$ modules later. A TP to DTP change therefore amounts to a real architecture change, and dense Transformer models need to be retrained or distilled after adaptation. We show that DTP breaks the permutation symmetry of neurons inside FFNs and of KV heads inside attention modules, and that this symmetry breakage makes the sharding itself a modelling decision. We show that maximising the affinity between the KV heads and the FFN neurons co-located on a device, by permuting the dense model before sharding, speeds up the distillation or retraining process. The affinity is measured with a first-order approximation of the damage that losing a head's contribution does to each neuron's output, and the co-located affinity is maximised with a coordinate-ascent optimiser that alternates an exact balanced assignment of neurons with an exhaustive search over the KV head partitions. The whole procedure takes under two minutes on one GPU for Qwen3-0.6B and Danube3-500M. On these models, at $\delta=1$, the affinity-optimised layouts reach any distillation target in about half to two thirds of the steps needed by the naive contiguous layouts, over the whole 10k-step range we tested, and every optimised seed beats every contiguous seed and all but one of the sixteen random layouts. We also show that the co-located affinity score at initialisation predicts the KL to the base model after training, across seventeen layouts ranging from anti-optimised to optimised (Pearson $-0.81$ and $-0.89$).

World models trained with joint-embedding predictive architectures learn compact, structured latent representations from physical interaction, yet planning in these latent spaces typically relies on one of two costly approaches. Search-based planners such as CEM, MPPI, and iCEM optimize action sequences through many predictor rollouts, achieving strong performance at the cost of high per-decision compute and latency. Policy-based methods amortize inference into a single forward pass but can degrade on contact-rich tasks where the demonstration distribution is multimodal. We propose LePlanner, an amortized iterative controller that learns to construct and refine latent action sequences through a frozen world-model predictor. LePlanner is trained with an arrival-and-hold objective that encourages the controller to reach the goal at the earliest feasible horizon and remain there. This addresses horizon-reset procrastination, a failure mode in which repeated receding-horizon replanning continually postpones goal arrival. An additional action-Gaussian loss keeps generated actions near the support of the offline dataset. Across navigation, contact-rich manipulation, and continuous-control environments, LePlanner matches or exceeds search-based planners while requiring an order of magnitude fewer predictor evaluations and 3-49x lower wall-clock time per decision. It achieves success rates of 98% on PushT, 100% on Reacher, 100% on TwoRooms, and 92% on the OGBench Cube task. These results show that much of the structure discovered through online search can be amortized into a lightweight iterative policy, enabling fast, horizon-aware, nonlinear physical control without online optimization.

This article investigates pre-training strategies for graph transformers in the biochemistry domain. By conducting comprehensive experiments, the study reveals that supervised pre-training using computed properties as labels provides the highest performance gain on downstream tasks. The results also highlight the importance of constraining model capacity to mitigate overfitting in graph transformers.

Recent advances in speech synthesis and voice conversion have made deepfake speech increasingly realistic, making generalization to unseen spoofing attacks a critical challenge. Pretrained speech and audio models offer a promising direction for improving robustness to such unseen attacks. Self-supervised learning (SSL) models capture fine-grained, low-level acoustic characteristics, whereas Auditory Large Language Models (ALLMs) provide higher-level contextual representations. These complementary views can provide useful cues for improving generalization to unseen attacks. However, direct fusion does not explicitly disentangle information shared across the two views from view-specific complementary information, limiting effective cross-view integration. To address this, we propose CRAF, a cross-view residual-aware fusion framework that uses ALLM-guided cross-view attention to enrich SSL representations and adopts ALLM as a high-level reference to separate ALLM-explainable information from complementary SSL residual information. The residual is selectively refined through adaptive gating and integrated through SSL-primary fusion. Experiments on ASVspoof 5 show that CRAF with Kimi-Audio achieves an EER of 5.96% and a minDCF of 0.1192, demonstrating robustness to unseen spoofing attacks.