arXiv:2605.17900v1 Announce Type: new
Abstract: Accurate Point of Interest (POI) attribute acquisition is essential for location-based services, yet traditional modular Interactive Voice Response (IVR) systems suffer from error accumulation and high maintenance overhead. We present DuIVRS-2, a large language model (LLM)-based end-to-end framework designed for large-scale POI attribute acquisition at Baidu Maps. To address the long-tail distribution of real-world interactions, our methodology first employs a finite state machine (FSM)-guided data augmentation strategy to synthesize a balanced and diverse training dataset. We then streamline dialogue management via a selective generation scheme combined with a Chain-of-Thought (CoT) mechanism, which ensures output stability and effectively eliminates hallucinations in industrial settings. To facilitate continuous policy refinement with minimal manual effort, we design a cooperative iterative learning framework that leverages a dual-evaluator voting system. Deployed in production for two months, DuIVRS-2 processed 0.4 million calls daily and achieved a 83.9\% Task Success Rate (TSR), outperforming its predecessor by 4 percentage points while maintaining a low reaction time of 130ms. This work provides a production-proven reference for developing robust, cost-effective LLM agents for large-scale industrial dialogue applications.
Science Journals
arXiv:2603.29167v2 Announce Type: replace
Abstract: We study a systems-level visual inference problem: using an expensive privileged modality during training while preserving a fixed-cost, single-modality deployment path. We present JDCNet, a confidence-gated CT-to-X-ray distillation framework in which the CT teacher supplies an auxiliary hard or temperature-scaled target only on training samples whose teacher confidence exceeds a threshold; at deployment the student takes X-ray input alone and matches the parameter, MAC, and latency profile of the supervised X-ray baseline. On a 510-patient same-patient paired BIMCV cohort with patient-level 5-fold cross-validation, two JDCNet configurations clear a fixed transfer gate against the supervised ResNet-18 baseline: 3-slice soft-KL supervision yields $\Delta\mathrm{BA}{=}{+}0.035$ ($95\%$ CI $[{+}0.011,{+}0.057]$) and mid-slice hard supervision yields $+0.033$ ($[{+}0.007,{+}0.058]$). Under the same splits and gate, logit distillation, gated logit distillation, contrastive alignment, attention transfer, feature hints, BiomedCLIP fine-tuning, and a module-augmented variant do not pass. Confidence-gated auxiliary targets are therefore a more transferable channel than uniformly softened CT logits; the evidence is bounded to one paired cohort, so external paired-cohort replication is required before any deployment claim.
arXiv:2605.18268v1 Announce Type: new
Abstract: Ion Cyclotron Emission (ICE) is a ubiquitous magnetised plasma phenomenon previously detected on virtually all large magnetic fusion devices and whose diagnostic potential for future power plants rests upon an accurate mapping of plasma parameters to spectra. This work demonstrates that the inverse problem is solved by training Time-Series Extrinsic Regression (TSER) models on synthetic ICE spectra from oblique propagation angle sweeps of nonlinear fully kinetic 1D3V particle-in-cell simulations of the magnetoacoustic cyclotron instability. Using datasets from a systematically constructed scan over reactor-relevant ranges of background magnetic field strength, density, and alpha-particle velocity pitch ($v_\parallel/|v|$) and concentration, we show that these bulk and fast ion parameters may be recovered from a JET ICE spectrum via TSER models with near real-time capability.
arXiv:2605.17716v1 Announce Type: new
Abstract: Sovereign network functions, e.g., routing protocols, are becoming increasingly complex and susceptible to failures arising from protocol configuration anomalies and anomalous configurations. This paper interprets the protocol configuration anomaly detection problem as detection of structural inconsistencies of connected nodes and edges in a bipartite graph that captures both physical network entities and logical protocol states. This graph structural inconsistency detector (GSID) model is proposed to solve the problem efficiently. To handle the heterogeneous nature of protocol configuration parameters, GSID employs an adaptive configuration encoder (ACE) that dynamically selects encoding strategies per parameter to preserve fine-grained numerical discrepancies. To expose the subtle inconsistencies of connected nodes and edges in the bipartite graph, GSID uses an inconsistency dynamic attention (IDA) mechanism that scores edges by drawing asymmetric attentions from both ends, rule compliance from one end and route connectivity from the other. It is demonstrated experimentally that GSID outperforms state-of-the-art baselines by threefold in F1 score and by 23.2% in accuracy. Ablation studies validate the effectiveness of both the ACE and IDA modules. Tests on unseen network scales and real-world network topologies show the superior adaptability of our GSID, compared to the baselines.
arXiv:2605.02832v2 Announce Type: replace
Abstract: Deciding how to distribute work between humans and AI systems is a central challenge in organisational design. Most approaches treat this as a binary choice, yet the operational reality is richer: humans and AI routinely share tasks or take complementary roles depending on context, fatigue, and the stakes involved. Governing that distribution -- balancing efficiency, oversight, and human capability -- remains an open problem. This paper presents Human-AI Adaptive Symbiosis (HAAS), an implemented framework for adaptive task allocation in software engineering and manufacturing. HAAS combines two coupled components: a rule-based expert system that enforces governance constraints before any learning occurs, and a contextual-bandit learner that selects among feasible collaboration modes from outcome feedback. Task-agent fit is represented through five auditable cognitive dimensions and a five-mode autonomy spectrum -- from human-only to fully autonomous -- embedded in a reproducible benchmark spanning both domains. Three empirical findings emerge. First, governance is not a binary switch but a tunable design variable: tighter constraints predictably convert autonomous AI assignments into supervised collaborations, with domain-specific costs and benefits. Second, in manufacturing, stronger governance can improve operational performance and reduce fatigue simultaneously -- a workload-buffering effect that contradicts the usual framing of governance as pure overhead. Third, no single governance setting dominates across all contexts; moderate governance becomes increasingly competitive as the learner accumulates experience within the governed action space. Together, these findings position HAAS as a pre-deployment workbench for comparing and inspecting human--AI allocation policies before organisational commitment.
arXiv:2602.13957v3 Announce Type: replace
Abstract: In this paper, we propose a data-enabled moving horizon estimation (MHE) approach for a class of nonlinear systems without explicit modeling, by leveraging Koopman operator theory and Willems fundamental lemma. Specifically, the nonlinear system is lifted to a linear parameter-varying Koopman surrogate, in which the lifting functions and scheduling mappings are learned directly from data using neural networks. Willems fundamental lemma is then employed to construct a trajectory-based representation of the Koopman surrogate, which bypasses the explicit identification of the matrices of the Koopman surrogate. Based on this representation, we formulate a convex data-enabled MHE design, which provides real-time estimates of the Koopman surrogate states, from which the states of the original nonlinear system are reconstructed. Sufficient conditions are derived to ensure the stability of the estimation error. The effectiveness of the proposed method is illustrated using a simulated membrane-based biological wastewater treatment process.
arXiv:2605.17712v1 Announce Type: new
Abstract: Generative Artificial Intelligence (GenAI) has prompted significant discussion in education, yet large-scale empirical evidence on how students and teachers perceive and navigate this shift remains limited. We analyse 270k AI-related Reddit posts and comments from 26 education-related subreddits spanning higher education, K-12 teaching, and professional training between November 2022 and April 2026. Topic modelling reveals seventeen themes covering academic integrity, teaching & pedagogy, career anxiety, policy, and niche professional contexts. Discourse evolves from an early detection-and-evasion arms race into a sustained enforcement regime that constructive integration only begins to challenge in mid-2024. Stakeholder communities differ sharply: K-12 teachers foreground cognitive dependency, academics focus on AI detection and deliberation, and professional-programme students concentrate on career anxiety. Sentiment correlates strongly negatively with engagement, showing adversarial enforcement themes mobilise communities far more than constructive integration discourse. Examining where faculty and students meet, we find 17% of threads are cross-role, and one third of such contact occurs in the adversarial themes AI Detection and Misconduct Enforcement. Students initiate 68% of mixed threads, but faculty produce most cross-role replies. Mixed threads contain 2-3 times more records and last 2-4 times longer than same-role threads, making adversarial integrity disputes the center of sustained faculty-student contact. We discuss implications for governance, pedagogical design, and cross-role contact design. The code and data is available at https://github.com/tugrulz/genai-edu
arXiv:2502.05462v2 Announce Type: replace
Abstract: We propose a real-time implementable motion planning framework for cooperative object transportation by nonholonomic mobile manipulator robots (MMRs) in dynamic environments. Our global planner finds a path from start to goal through the static, obstacle-free regions in the environment and generates a set of convex, static, obstacle-free regions around the path using a novel, fast, and computationally lightweight ellipse-based technique. We introduce a nonlinear Model Predictive Control (NMPC) based real-time implementable planning technique that jointly plans feasible motion for the mobile base and the manipulator's arm and generates a kinodynamic feasible, collision-free trajectory for cooperative object transportation. Simulation and hardware experiments validate the efficiency of our proposed planning framework.
arXiv:2407.17130v2 Announce Type: replace
Abstract: The mathematical formulation of sign-changing problems involves a linear second-order partial differential equation in the divergence form, where the coefficient can assume positive and negative values in different subdomains. These problems find their physical background in negative-index metamaterials, either as inclusions embedded into common materials as the matrix or vice versa. In this paper, we propose a numerical method based on the constraint energy minimizing generalized multiscale finite element method (CEM-GMsFEM) specifically designed for sign-changing problems. The construction of auxiliary spaces in the original CEM-GMsFEM is tailored to accommodate the sign-changing setting. The numerical results demonstrate the effectiveness of the proposed method in handling sophisticated coefficient profiles and the robustness of coefficient contrast ratios. Under several technical assumptions and by applying the \texttt{T}-coercivity theory, we establish the inf-sup stability and provide an a priori error estimate for the proposed method.
arXiv:2605.08738v2 Announce Type: replace
Abstract: Structured pruning and knowledge distillation (KD) are typical techniques for compressing large language models, but it remains unclear how they should be applied at pretraining scale, especially to recent mixture-of-experts (MoE) models. In this work, we systematically study MoE compression in large-scale pretraining, focusing on three key questions: whether pruning provides a better initialization than training from scratch, how expert compression choices affect the final model after continued training, and which training strategy is most effective. We have the following findings: First, across depth, width, and expert compression, pruning a pretrained MoE consistently outperforms training the target architecture from scratch under the same training budget. Second, different one-shot expert compression methods converge to similar final performance after large-scale continual pretraining. Motivated by this, we introduce a simple partial-preservation expert merging strategy that improves downstream performance across most benchmarks. Third, combining KD with the language modeling loss outperforms KD alone, particularly on knowledge-intensive tasks. We further propose multi-token prediction (MTP) distillation, which yields consistent gains. Finally, given the same training tokens, progressive pruning schedules outperform one-shot compression, suggesting that gradual architecture transitions lead to better optimization trajectories. Putting it all together, we compress Qwen3-Next-80A3B to a 23A2B model that retains competitive performance. These results offer practical guidance for efficient MoE compression at scale.
arXiv:2407.03175v2 Announce Type: replace
Abstract: This note demonstrates that we can stably recover all symmetric Toeplitz matrices $\pmb{X}_0\in\mathbb{R}^{n\times n}$ of rank at most $r$ from a number of rank-one subgaussian measurements on the order of $r\log^{2} n$ with an exponentially decreasing failure probability by employing a nuclear norm minimization program. Our approach utilizes descent cone analysis through Mendelson's small ball method with the Toeplitz constraint. The key ingredient is to determine the spectral norm of a random matrix with Toeplitz structure, which may be of independent interest. This improves upon earlier analyses and resolves the conjecture in Chen et al. (IEEE Transactions on Information Theory, 61(7):4034--4059, 2015).
arXiv:2605.18482v1 Announce Type: new
Abstract: The growing adoption of lattice-based structures in soft robotics creates a need for advanced sensing solutions capable of monitoring their global deformation, particularly compression and extension. In this work, we address this challenge by introducing a novel optical sensor based on two patterned waveguides arranged in an ellipsoidal geometry. This Bidirectional Optical sensor for Actuation Tracking (BOAT) is seamlessly co-printed with a lattice structure actuated by an embedded pneumatic artificial muscle (PAM), and its performance is assessed. During PAM elongation or contraction, the bending of the embedded BOAT waveguides induces output signal variations that enable a clear discrimination between compression and extension states.
The designs of both each specific waveguide structure (by surface patterning) and of the sensorized lattice-based unit embedding two BOATs are supported by numerical simulations. Experimental calibration over 100 consecutive pressure cycles ranging from +50 kPa to $-$40 kPa demonstrates a highly repeatable response, allowing a reliable distinction between extension and compression.
Finally, sensor feedback is used to implement a digital shadow, enabling continuous synchronization between the whole sensorized unit and its virtual counterpart. These results establish BOAT as a powerful and reliable approach for deformation monitoring in soft lattice-based robotic systems.
arXiv:2603.16091v3 Announce Type: replace
Abstract: In factual question answering, many errors are not failures of access but failures of commitment: the system retrieves relevant evidence, yet still settles on the wrong answer. We present CounterRefine, a lightweight repair layer for short-form RAG that treats the first answer as a hypothesis to test. Given a draft, CounterRefine issues answer-conditioned expansion queries to retrieve candidate-specific evidence, then applies a constrained KEEP or REVISE refinement step whose proposed revisions are accepted only after deterministic validation. The design is intentionally narrow: it adds one evidence-gathering pass and one guarded refinement call rather than replacing the retriever or building a broad agentic system. On the full SimpleQA benchmark, CounterRefine improves a matched one-pass RAG baseline by up to 5.8 correct-rate points; in the full Claude trace, it changes only 5.6% of outputs, with 180 beneficial outcome changes and 8 harmful ones. These findings suggest a simple but important direction for knowledgeable foundation models: beyond accessing evidence, they should also be able to use that evidence to reconsider and, when necessary, repair their own answers.
arXiv:2605.16898v1 Announce Type: cross
Abstract: During the propagation of cosmic rays in the solar system, the Sun will block those particles and form a shadow whose position and depth are very important probe of the magnetic fields in the Sun's corona, in the interplanetary space, and the Earth's vicinity. In this work we carry out Monte Carlo studies of the Sun shadow, with a novel approach to take into account daily variations of the coronal and interplanetary magnetic field models. This treatment is suitable for studies of short-term variations of the Sun shadow, which become detectable by the Large High Altitude Air Shower Observatory (LHAASO) experiment. Two different coronal magnetic field models, the Potential Field Source Surface (PFSS) and Current Sheet Source Surface (CSSS) models, with observational time-varying photospheric magnetic fields as boundary conditions, are studied in this work. The interplanetary magnetic fields are then derived using the Parker spiral model based on the coronal ones. Furthermore, both the coronal and interplanetary magnetic field strengths are corrected using the Parker Solar Probe (PSP) measurements. We compare the simulation results with the daily observations of Sun shadow by LHAASO in 2021, and find that the CSSS model generally shows better consistency of the displacement of the Sun shadow than the PFSS model.
arXiv:2204.01611v3 Announce Type: replace
Abstract: Inspired by the cognitive science theory, we explicitly model an agent with both semantic and episodic memory systems, and show that it is better than having just one of the two memory systems. In order to show this, we have designed and released our own challenging environment, "the Room", compatible with OpenAI Gym, where an agent has to properly learn how to encode, store, and retrieve memories to maximize its rewards. The Room environment allows for a hybrid intelligence setup where machines and humans can collaborate. We show that two agents collaborating with each other results in better performance than one agent acting alone.
arXiv:2605.18484v1 Announce Type: new
Abstract: The rapid adoption of Web3 infrastructures has led to a growing number of security incidents affecting cryptocurrency exchanges, custody services and blockchain-based platforms. While existing research predominantly focuses on vulnerabilities in smart contracts and blockchain protocols, a substantial portion of real-world losses originates from off-chain systems, organizational processes and human-centered operational workflows.
This paper presents a qualitative, incident-based analysis of publicly documented, high-impact security breaches in the Web3 ecosystem, including the Bybit exchange incident (2025), the Ronin Network bridge compromise (2022), and the DMM Bitcoin exchange breach (2024). The selected cases are systematically analysed and mapped to established Web2 security reference frameworks, including OWASP-based vulnerability categories and organizational security control domains.
The results indicate that dominant failure patterns in Web3 environments are insufficiently addressed by generic security control catalogues, particularly with respect to cryptographic key management, transaction approval governance, signer and validator infrastructure, third-party tooling dependencies, and human-in-the-loop processes.
Based on these findings, this paper argues for the adoption of established information security management systems (ISMS) in Web3 organizations and derives a structured set of blockchain-specific cybersecurity control categories to operationalize existing ISMS frameworks for blockchain-based systems. The proposed categories aim to bridge the gap between generic security governance frameworks and domain-specific risks inherent to Web3 infrastructures.
arXiv:2605.01235v2 Announce Type: replace
Abstract: Driven by the escalating global burden of mental health conditions, music-based interventions have attracted significant attention as a non-invasive, cost-effective modality for emotion regulation and psychological stress relief. However, current digital music services rely on static preferences and fail to adapt to users' instantaneous psychological states. Furthermore, directly mapping electroencephalography (EEG) to music generation remains challenging due to severe paired-data scarcity and a lack of interpretability. To address these limitations, we propose MindMelody, a fully functional, closed-loop real-time system for EEG-driven personalized music intervention. MindMelody introduces an emotion-mediated semantic bridge. Specifically, a hybrid Transformer-GNN first decodes real-time EEG signals into global Valence-Arousal states and local temporal affect trajectories. These states are then fed into a Retrieval-Augmented Generation (RAG)-equipped Large Language Model (LLM) to formulate structured intervention plans. Subsequently, a novel Hierarchical EEG Controller injects global affect prefixes and local temporal guidance into a pretrained music backbone, enabling fine-grained controllable audio synthesis. Crucially, the system incorporates a continuous feedback loop that updates generation parameters on the fly based on the user's evolving EEG dynamics. Extensive experiments show that MindMelody improves control adherence and emotional alignment, and receives higher perceived helpfulness in a short-term listening setting, suggesting its promise as an adaptive affect-aware music generation framework.
arXiv:2512.04331v2 Announce Type: replace
Abstract: The surge in face forgeries has increasingly undermined confidence in the authenticity of online content. As generation algorithms rapidly evolve, new fake categories will constantly emerge, severely challenging existing face forgery detection methods. Although face forgery detection has recently improved, current techniques remain largely confined to binary Real-vs-Fake classification or the recognition of known fake categories. Moreover, they fail to identify the emergence of entirely new forgery methods. In this work, we study the Open Set Face Forgery Detection (OSFFD) problem, which requires the detection model to identify novel fake categories. To enhance its real-world applicability, we reformulate the OSFFD problem and address it through uncertainty estimation. Specifically, we propose the Dual-Level Evidential face forgery Detection (DLED) approach, which estimates prediction uncertainty by extracting and integrating category-specific evidence on the spatial and frequency levels. Comprehensive experiments across diverse settings demonstrate that our proposed DLED approach achieves state-of-the-art performance. Notably, it surpasses various existing baseline models by a $20\%$ margin on average when identifying forgeries from novel fake categories. Concurrently, our DLED method yields competitive performance on the standard binary Real-versus-Fake face forgery detection task.
arXiv:2605.18511v1 Announce Type: new
Abstract: A lightweight and reproducible denoising pipeline for high-throughput Raman spectroscopy is presented. The approach relies on a one-dimensional convolutional autoencoder trained using a Noise2Noise strategy, requiring neither external spectral libraries nor high signal-to-noise reference spectra for training. From a reduced training subset composed of repeated short-exposure acquisitions, the model learns to reconstruct Raman spectra while efficiently suppressing stochastic noise. The method is evaluated on a heterogeneous mineral sample, using both quantitative spectral fidelity metrics (RMSE, SNR, SSIM) and task-oriented criteria based on unsupervised K-means classification. Results demonstrate that integration times as short as 5 ms per spectrum, which are typically insufficient for reliable interpretation, yield denoised spectra with high fidelity to the reference data while preserving chemically coherent maps. This work provides a practical trade-off between spectral quality and acquisition speed, enabling fast, adaptable Raman workflows compatible with routine laboratory use. It also offers a transferable framework for other one-dimensional spectroscopic modalities.
arXiv:2605.18309v1 Announce Type: new
Abstract: Although Large Language Models (LLMs) achieve strong alignment through supervised fine-tuning and reinforcement learning from human feedback, the alignment is often fragile under subsequent fine-tuning. Existing explanations either attribute alignment fragility to gradient geometry or characterize it as a distributional shift in model outputs, yet few provide a unified account that bridges parameter-space learning dynamics with function-space alignment behavior during fine-tuning. In this work, we introduce a tractable alignment score and derive its closed-form update during fine-tuning, yielding a unified framework for alignment dynamics. Our analysis decomposes alignment updates into two competing components: a \textbf{\color{red!60!black} Rebound Force}, governed jointly by the current alignment state and the narrowness of model distribution, and a \textbf{\color{green!60!black} Driving Force}, determined by how the training distribution aligns with outcome-conditioned posteriors over aligned and non-aligned completions. This decomposition explains why prior alignment can be reversed by later fine-tuning and why narrower posterior structure strengthens such reversal. Moreover, our framework predicts a \textbf{Rehearsal Priming Effect}: prior alignment leaves a latent posterior imprint that amplifies the effective Driving Force upon re-exposure, leading to faster re-alignment. We validate these predictions across safety alignment, emergent misalignment, and sentiment settings, demonstrating consistent alignment reversal and accelerated re-alignment under re-exposure. In addition, controlled experiments in safety alignment confirm the predicted dependence of rebound strength on posterior narrowness. Together, these results provide a unified dynamical perspective on how alignment is disrupted and reactivated during LLM fine-tuning.
arXiv:2605.18719v1 Announce Type: new
Abstract: Diffusion models have been widely studied for removing unsafe content learned during pre-training. Existing methods require expensive supervised data, either unsafe-text paired with safe-image groundtruth or negative/positive image pairs, making them impractical to scale. Furthermore, offline reinforcement learning and supervised fine-tuning approaches that generate synthetic data offline suffer from catastrophic forgetting, degrading generation quality. We propose a novel online reinforcement learning framework that addresses both data scarcity and model degradation through post-training with Group Relative Policy Optimization (GRPO) on both negative and positive text prompts. To eliminate the need for fine-tuning specialized safe/unsafe reward models, we introduce a \textit{steering reward mechanism} that exploits an inherent property of CLIP embeddings: steering text representations toward positive safety directions and away from negative ones in the embedding space. Our online-policy approach enables the model to learn from diverse prompts, including explicit unsafe content, without catastrophic forgetting. Extensive experiments demonstrate that our method reduces inappropriate content to 18.07\% (vs. 48.9\% for SD v1.4) and nudity detections to 15 (vs. 646 baseline) while improving compositional generation quality from 42.08\% to 47.83\% on GenEval. Remarkably, these safety gains generalize to out-of-domain unsafe prompts across seven harm categories, achieving state-of-the-art performance without supervised paired data or reward tuning. Github: https://github.com/MAXNORM8650/SafeDiffusion-R1.
arXiv:2605.18345v1 Announce Type: cross
Abstract: Hybrid quantum-classical neural networks (HQNNs) are emerging as a practical approach for quantum machine learning in the noisy intermediate-scale quantum (NISQ) era, as they combine classical learning components with parameterized quantum circuits in an end-to-end trainable framework. However, their performance and efficiency depend strongly on architectural choices such as data encoding, circuit structure, measurement design, and the coupling between classical and quantum modules. This makes manual design increasingly difficult, especially when hardware limitations and resource constraints must also be taken into account. In this paper, we study the foundations of HQNNs and neural architecture search (NAS), discuss how NAS extends to quantum and hybrid settings, and demonstrate FLOPs-aware search (where FLOPs serve as a proxy for computational complexity), as an important hardware-aware direction for building HQNNs that are not only accurate but also computationally efficient and practically deployable.
arXiv:2601.02353v3 Announce Type: replace
Abstract: Farmers in remote areas need quick and reliable methods for identifying plant diseases, yet they often lack access to laboratories or high-performance computing resources. Deep learning models can detect diseases from leaf images with high accuracy, but these models are typically too large and computationally expensive to run on low-cost edge devices such as Raspberry Pi. Furthermore, collecting thousands of labeled disease images for training is both expensive and time-consuming. This paper addresses both challenges by combining neural network pruning, removing unnecessary parts of the model, with few-shot learning, which enables the model to learn from limited examples. This paper proposes Disease-Aware Channel Importance Scoring (DACIS), a method that identifies which parts of the neural network are most important for distinguishing between different plant diseases, integrated into a three-stage Prune-then-Meta-Learn-then-Prune (PMP) pipeline. Experiments on PlantVillage and PlantDoc datasets demonstrate that the proposed approach reduces model size by 78% while maintaining 92.3% of the original accuracy, with the compressed model running at 7 frames per second on a Raspberry Pi 4, making real-time field diagnosis practical for smallholder farmers.
arXiv:2601.09413v2 Announce Type: replace
Abstract: We introduce a voice-agentic framework that learns one critical omni-understanding skill: knowing when to trust itself versus when to consult external audio perception. Our work is motivated by a crucial yet counterintuitive finding: naively fine-tuning an omni-model on both speech recognition and external sound understanding tasks often degrades performance, as the model can be easily misled by noisy hypotheses. To address this, our framework, Speech-Hands, recasts the problem as an explicit self-reflection decision. This learnable reflection primitive proves effective in preventing the model from being derailed by flawed external candidates. We show that this agentic action mechanism generalizes naturally from speech recognition to complex, multiple-choice audio reasoning. Across the OpenASR leaderboard, Speech-Hands consistently outperforms strong baselines by 12.1% WER on seven benchmarks. The model also achieves 77.37% accuracy and high F1 on audio QA decisions, showing robust generalization and reliability across diverse audio question answering datasets. By unifying perception and decision-making, our work offers a practical path toward more reliable and resilient audio intelligence.
arXiv:2604.02178v2 Announce Type: replace
Abstract: Mixture-of-Experts (MoE) architectures have become the dominant choice for scaling Large Language Models (LLMs), activating only a subset of parameters per token. While MoE architectures are primarily adopted for computational efficiency, it remains an open question whether their sparsity makes them inherently easier to interpret than dense feed-forward networks (FFNs). We compare MoE experts and dense FFNs using $k$-sparse probing and find that expert neurons are consistently less polysemantic, with the gap widening as routing becomes sparser. This suggests that sparsity pressures both individual neurons and entire experts toward monosemanticity. Leveraging this finding, we zoom out from the neuron to the expert level as a more effective unit of analysis. We validate this approach by automatically interpreting hundreds of experts. This analysis allows us to resolve the debate on specialization: experts are neither broad domain specialists (e.g., biology) nor simple token-level processors. Instead, they function as fine-grained task experts, specializing in linguistic operations or semantic tasks (e.g., closing brackets in $\LaTeX{}$). Our findings suggest that MoEs are inherently interpretable at the expert level, providing a clearer path toward large-scale model interpretability. Code is available at: https://github.com/jerryy33/MoE_analysis.