Interactively learning from observation and language feedback is an increasingly studied area driven by the emergence of large language model (LLM) agents. Despite impressive empirical demonstrations, so far a principled framing of these decision problems remains lacking. We formalize the Learning from Language Feedback (LLF) problem, assert sufficient assumptions to enable learning despite latent rewards, and introduce $\textit{transfer eluder dimension}$ as a measure to characterize the hardness of LLF. We formalize the intuition that information in the language feedback governs the learning complexity, and demonstrate cases where learning from rich language feedback can be exponentially faster than learning from reward. We develop a no-regret algorithm, called $\texttt{HELiX}$, that provably solves LLF problems through sequential interactions, with performance guarantees that scale with the transfer eluder dimension. Across several empirical domains, we show that $\texttt{HELiX}$ performs well even when repeatedly prompting LLMs does not work reliably. Our contributions mark an important step towards designing principled interactive learning algorithms using generic language feedback.

Autoregressive Models (ARMs) have long dominated the landscape of Large Language Models. Recently, a new paradigm has emerged in the form of diffusion-based Large Language Models (dLLMs), which generate text by iteratively denoising masked segments. This approach has shown significant advantages and potential. However, dLLMs suffer from high inference latency. Traditional ARM acceleration techniques, such as Key-Value caching, are incompatible with dLLMs due to their bidirectional attention mechanism. To address this specific challenge, our work begins with a key observation that dLLM inference involves a static prompt and a partially dynamic response, where most tokens remain stable across adjacent denoising steps. Based on this, we propose dLLM-Cache, a training-free adaptive caching framework that combines long-interval prompt caching with partial response updates guided by feature similarity. This design enables efficient reuse of intermediate computations without compromising model performance. Extensive experiments on representative dLLMs, including LLaDA 8B and Dream 7B, show that dLLM-Cache achieves up to 9.1x FLOPs reduction on LongBench-HotpotQA while maintaining competitive output quality. Notably, our method brings dLLM inference latency close to that of ARMs under many settings. The code for this work is publicly available at: https://github.com/maomaocun/dLLM-cache.

Active vision, also known as active perception, refers to actively selecting where and how to look in order to gather task-relevant information. It is a critical component of efficient perception and decision-making in humans and advanced embodied agents. With the rise of Multimodal Large Language Models (MLLMs) as central planners in robotic systems, the lack of methods for equipping MLLMs with active perception has become a key gap. We first provide a systematic definition of MLLM-based active perception tasks and show that GPT-o3's zoom-in strategy can be viewed as a special case, though it suffers from low efficiency and inaccurate region selection. To address these issues, we propose ACTIVE-o3, a reinforcement learning framework built on GRPO that equips MLLMs with active perception capabilities. Leveraging a modular sensing-action design and a dual-form reward, ACTIVE-o3 autonomously learns efficient and stable region selection strategies without explicit region-selection supervision. We further establish a comprehensive benchmark covering both open-world tasks, including small- and dense-object grounding, and domain-specific scenarios, including remote sensing, autonomous driving, and interactive segmentation. Experimental results demonstrate that ACTIVE-o3 significantly enhances active perception capabilities compared to baselines. Moreover, we show that our framework not only preserves the model's general understanding ability but can also serve as a proxy task for leveraging perception data, further improving performance on benchmarks such as RealWorldQA and MME.

Cognitive Diagnosis has become a critical task in AI-empowered education, supporting personalized learning by accurately assessing students' cognitive states. However, traditional cognitive diagnosis models (CDMs) often struggle in cold-start scenarios due to the lack of student-exercise interaction data. Recent NLP-based approaches leveraging pre-trained language models (PLMs) have shown promise by utilizing textual features, but they fail to fully bridge the gap between semantic understanding and cognitive profiling. To address this limitation, we propose \textbf{L}anguage \textbf{M}odel-based \textbf{C}ognitive \textbf{D}iagnosis (LMCD), a unified, LLM-adaptable framework designed to tackle cold-start challenges by harnessing the advanced capabilities of large language models (LLMs). LMCD operates via two primary phases: (1) Knowledge Diffusion, where LLMs generate enriched content for exercises and knowledge concepts (KCs) to establish stronger semantic links; and (2) Semantic-Cognitive Fusion, which leverages LLMs to deeply integrate textual information with student cognitive states. By unifying the semantic and cognitive spaces, LMCD creates comprehensive representations that serve as a plug-and-play enhancement for various off-the-shelf CDMs. Experiments on two real-world datasets demonstrate that LMCD significantly outperforms state-of-the-art methods in both exercise-cold and domain-cold settings. https://github.com/TAL-auroraX/LMCDThe code is publicly available at https://github.com/TAL-auroraX/LMCD

We present the first provable method for identifying symmetric linear dynamical systems (LDS) with accuracy guarantees that are independent of the systems' state dimension or effective memory. Our approach builds upon recent work that represents symmetric LDSs as convolutions learnable via fixed spectral transformations. We show how to invert this representation, thereby recovering an LDS model from its spectral transform and yielding an end-to-end convex optimization procedure. This distillation preserves predictive accuracy while enabling constant-time and constant-space inference per token, independent of sequence length. We evaluate our method, SpectraLDS, as a component in sequence prediction architectures and demonstrate that accuracy is preserved while inference efficiency is improved on tasks such as language modeling.

Neural network pruning aims to compress models for efficient deployment, yet two fundamental challenges remain. First, many methods rely on per-component importance scores, selecting filters or neurons independently and ignoring redundancy: the retained set may include multiple components capturing similar discriminative patterns while missing others entirely. Second, determining per-layer pruning ratios typically requires manual, architecture-specific tuning with no principled stopping criterion. We propose CoSeP (Complementary Separability Pruning) to address both issues. Rather than scoring components in isolation, CoSeP represents each component by its class-separability profile across all class pairs, computed via Jeffries--Matusita distances. This defines a separability space in which nearby components are potentially redundant and distant components capture complementary information. CoSeP selects a compact set of representatives in this space: components are grouped via k-medoids clustering, candidate subset sizes are evaluated using the Mean Simplified Silhouette, and a knee-detection criterion automatically determines how many components to retain. Across CIFAR-10, CIFAR-100, and ImageNet-1K, on ResNet, VGG, MobileNet, and DenseNet architectures, CoSeP matches or improves accuracy while reducing FLOPs, with measured wall-clock inference-time reductions of up to 20%. For example, it achieves a +0.66% top-1 accuracy gain with 2.30x FLOPs reduction on ResNet-50/ImageNet-1K, and a 0.37% gain with 2.59x FLOPs reduction on VGG-16/CIFAR-10. These results demonstrate that modeling complementarity in class-separability space provides an effective and principled approach to pruning.

Renal mass segmentation has important potential to enhance the clinical workflow, especially in settings requiring quantitative assessments. Kidney volume could serve as an important biomarker for renal diseases, with changes in volume correlating directly with kidney function. Currently, clinical practice often relies on subjective visual assessment for evaluating kidney size and kidney lesions, including tumors and cysts, which are typically staged based on diameter, volume, and anatomical location. To support a more objective and reproducible approach, this research aims to develop a robust, thoroughly validated renal mass segmentation algorithm, named Renal-Net. We employ publicly available training datasets and leverage the state-of-the-art medical image segmentation framework nnU-Net. Validation is conducted using both proprietary and public test datasets, with segmentation performance quantified by Dice coefficient and the 95th percentile Hausdorff distance. Furthermore, we analyze robustness across subgroups based on patient sex, age, CT contrast phases, and tumor histologic subtypes. Our findings demonstrate that our segmentation algorithm, trained exclusively on publicly available data, generalizes effectively to external test sets and outperforms existing state-of-the-art models across all tested datasets. Subgroup analyses reveal consistent high performance, indicating strong robustness and reliability. The developed algorithm and associated code are publicly accessible at https://github.com/DIAGNijmegen/oncology-kidney-abnormality-segmentation.

Cooperative perception, leveraging shared information from multiple vehicles via vehicle-to-vehicle (V2V) communication, plays a vital role in autonomous driving to alleviate the limitation of single-vehicle perception. Existing works have explored the effects of V2V communication impairments on perception precision, but they lack generalization to different levels of impairments. In this work, we propose a joint weighting and denoising framework, Coop-WD, to enhance cooperative perception subject to V2V channel impairments. In this framework, the self-supervised contrastive model and the conditional diffusion probabilistic model are adopted hierarchically for vehicle-level and pixel-level feature enhancement. An efficient variant model, Coop-WD-eco, is proposed to selectively deactivate denoising to reduce processing overhead. Rician fading, non-stationarity, and time-varying distortion are considered. Simulation results demonstrate that the proposed Coop-WD outperforms conventional benchmarks in all types of channels. Qualitative analysis with visual examples further proves the superiority of our proposed method. The proposed Coop-WD-eco achieves up to 50% reduction in computational cost under severe distortion while maintaining comparable accuracy as channel conditions improve.

Rapid global population growth underscores the need for digitally enabled agricultural systems that support sustainable food production and data-driven resource management for farmers and stakeholders. The adoption of Internet of Things (IoT) technologies, capable of capturing real-time environmental (e.g., temperature, humidity) and operational (e.g., irrigation) parameters, is a crucial step toward enabling advanced applications such as AI-based yield forecasting. However, the effectiveness of such models is often constrained by limited data availability, particularly in dynamic farm environments where IoT observations must be accumulated over multiple growing seasons. In this study, we deployed IoT sensors in strawberry production polytunnels over two growing seasons to collect data on water usage, internal and external temperature and humidity, soil moisture, soil temperature, and photosynthetically active radiation. These observations were combined with manually recorded yield data spanning four seasons. To address gaps in IoT data for the two seasons without sensor coverage, we developed an AI-based backcasting approach that synthesizes missing sensor observations using historical weather data from a nearby station and existing polytunnel measurements. We then trained AI-based yield forecasting models using both real and synthetic datasets. In this retrospective evaluation, results show that incorporating synthetic data improved yield forecasting accuracy, with models trained on the combined dataset outperforming those using only real sensor, weather, and yield data.

Chinese Grammatical Error Correction (CGEC) is a critical task in Natural Language Processing, addressing the growing demand for automated writing assistance in both second-language (L2) and native (L1) Chinese writing. While L2 learners struggle with mastering complex grammatical structures, L1 users also benefit from CGEC in academic, professional, and formal contexts where writing precision is essential. This survey provides a comprehensive review of CGEC research, covering datasets, annotation schemes, evaluation methodologies, and system advancements. We examine widely used CGEC datasets, highlighting their characteristics, limitations, and the need for improved standardization. We also analyze error annotation frameworks, discussing challenges such as word segmentation ambiguity and the classification of Chinese-specific error types. Furthermore, we review evaluation metrics, focusing on their adaptation from English GEC to Chinese, including character-level scoring and the use of multiple references. In terms of system development, we trace the evolution from rule-based and statistical approaches to neural architectures, including Transformer-based models and the integration of large pre-trained language models. By consolidating existing research and identifying key challenges, this survey provides insights into the current state of CGEC and outlines future directions, including refining annotation standards to address segmentation challenges, and leveraging multilingual approaches to enhance CGEC.

The Segment Anything Model 2 (SAM2), a prompt-guided video foundation model, has remarkably performed in video object segmentation, drawing significant attention in the community. Due to the high similarity between camouflaged objects and their surroundings, which makes them difficult to distinguish even by the human eye, the application of SAM2 for automated segmentation in real-world scenarios faces challenges in camouflage perception and reliable prompts generation. To address these issues, we propose CamoSAM2, a motion-appearance prompt inducer (MAPI) and refinement framework to automatically generate and refine prompts for SAM2, enabling high-quality automatic detection and segmentation in VCOD task. Initially, we introduce a prompt inducer that simultaneously integrates motion and appearance cues to detect camouflaged objects, delivering more accurate initial predictions than existing methods. Subsequently, we propose a video-based adaptive multi-prompts refinement (AMPR) strategy tailored for SAM2, aimed at mitigating prompt error in initial coarse masks and further producing good prompts. Specifically, we introduce a novel three-step process to generate reliable prompts by camouflaged object determination, pivotal prompt frame selection, and multi-prompts formation. Extensive experiments conducted on two benchmark datasets demonstrate that our proposed model, CamoSAM2, significantly outperforms existing state-of-the-art methods, achieving increases of 8.0% and 10.1% in mIoU metric. Additionally, our method achieves the fastest inference speed compared to current VCOD models.

Learning to transform conditional probability densities over time is a fundamental challenge spanning probabilistic modeling and the natural sciences. This task is paramount when forecasting the evolution of stochastic nonlinear dynamical systems in biological and physical domains. While flow-based models can predict the temporal evolution of probability distributions, existing approaches often assume discrete conditioning with samples that are paired across time, limiting their scientific applicability where frequently only sparse data with unpaired continuous conditioning is available. We propose Conditional Variable Flow Matching (CVFM), a framework for learning flows transforming conditional distributions with amortization across the continuous space of conditional densities. CVFM addresses the high-variance instability of prior methods by jointly sampling flows over state and conditioning variables, utilizing a conditioning mismatch kernel alongside a conditional Wasserstein distance to reweight the conditional optimal transport objective. Collectively, these advances allow for learning dynamics from sparse unpaired measurements of state-condition across time. We evaluate CVFM on conditional mapping benchmarks and a case study modeling the temporal evolution of materials internal structure during manufacturing processes, observing improved performance and convergence characteristics over existing conditional variants. Code is available at https://github.com/agenerale/conditional-variable-flow-matching.

With the rapid advancement of neural audio codecs, codec-based speech generation (CoSG) systems have become highly powerful. Unfortunately, CoSG also enables the creation of highly realistic deepfake speech, making it easier to mimic an individual's voice and spread misinformation. We refer to this emerging deepfake speech generated by CoSG systems as CodecFake. Detecting such CodecFake is an urgent challenge, yet most existing systems primarily focus on detecting fake speech generated by traditional speech synthesis models. In this paper, we introduce CodecFake+, a large-scale dataset designed to advance CodecFake detection. To our knowledge, CodecFake+ is the largest dataset encompassing the most diverse range of codec architectures. The training set is generated through re-synthesis using 31 publicly available open-source codec models, while the evaluation set includes web-sourced data from 17 advanced CoSG models. We also propose a comprehensive taxonomy that categorizes codecs by their root components: vector quantizer, auxiliary objectives, and decoder types. Our proposed dataset and taxonomy enable detailed analysis at multiple levels to discern the key factors for successful CodecFake detection. At the individual codec level, we validate the effectiveness of using codec re-synthesized speech (CoRS) as training data for large-scale CodecFake detection. At the taxonomy level, we show that detection performance is strongest when the re-synthesis model incorporates disentanglement auxiliary objectives or a frequency-domain decoder. Furthermore, from the perspective of using all the CoRS training data, we show that our proposed taxonomy can be used to select better training data for improving detection performance. Overall, we envision that CodecFake+ will be a valuable resource for both general and fine-grained exploration to develop better anti-spoofing models against CodecFake.

Applying machine learning to increasingly high-dimensional problems with sparse or biased training data increases the risk that a model is used on inputs outside its training domain. For such out-of-distribution (OOD) inputs, the model can no longer make valid predictions, and its error is potentially unbounded. Since testing OOD detection methods on real-world datasets is complicated, we design a benchmark for OOD detection, which includes three novel and easily-visualisable toy examples. These simple examples provide direct and intuitive insight into whether the detector is able to detect (1) linear and (2) non-linear concepts and (3) identify thin in-distribution (ID) subspaces (needles) within high-dimensional spaces (haystacks). We use our benchmark to evaluate the performance of various methods from the literature. Since tactile examples of OOD inputs may benefit OOD detection, we also review several simple methods to synthesise OOD inputs for supervised training. We introduce two improvements, $t$-poking and OOD sample weighting, to make supervised detectors more precise at the ID-OOD boundary. This is especially important when conflicts between real ID and synthetic OOD sample blur the decision boundary. Finally, we provide recommendations for constructing and applying OOD detectors in machine learning.

Recent advancements in audio tokenization have significantly enhanced the integration of audio capabilities into large language models (LLMs). However, audio understanding and generation are often treated as distinct tasks, hindering the development of truly unified audio-language models. While instruction tuning has demonstrated remarkable success in improving generalization and zero-shot learning across text and vision, its application to audio remains largely unexplored. A major obstacle is the lack of comprehensive datasets that unify audio understanding and generation. To address this, we introduce Audio-FLAN, a large-scale instruction-tuning dataset covering 80 diverse tasks across speech, music, and sound domains, with over 100 million instances. Audio-FLAN lays the foundation for unified audio-language models that can seamlessly handle both understanding (e.g., transcription, comprehension) and generation (e.g., speech, music, sound) tasks across a wide range of audio domains in a zero-shot manner. The Audio-FLAN dataset is available on HuggingFace and GitHub.

This paper presents a framework for generating 3D indoor scenes from text prompts. Existing methods often formulate scene synthesis as an object layout prediction problem conditioned on a single input modality, such as a text description, room shape, or scene graph. This design can lead to object collisions and limited functional plausibility, reducing its practical applicability. To address these limitations, we introduce a multi-stage pipeline that better reflects practical scene creation scenarios. Given a text prompt describing partial scene content, our method first uses graph diffusion to produce a contextually coherent scene graph and then predicts a realistic object layout. In addition, we incorporate lightweight human-object interaction priors to encourage human-centric and functional arrangements, with explicit spatial constraints to reduce interpenetration. Our approach generates coherent 3D scenes with viable layouts that better support human interaction. Experiments on the 3D-FRONT dataset demonstrate that our method achieves competitive or state-of-the-art performance compared with existing approaches, while improving the physical plausibility of generated scenes.

We investigate diffusion models to solve the Traveling Salesman Problem. Building on the recent DIFUSCO and T2TCO approaches, we propose IDEQ. IDEQ improves the quality of the solutions by leveraging the constrained structure of the state space of the TSP. Another key component of IDEQ consists in replacing the last stages of DIFUSCO curriculum learning by considering a uniform distribution over the Hamiltonian tours whose orbits by the 2-opt operator converge to the optimal solution as the training objective. Our experiments show that IDEQ improves the state of the art for such neural network based techniques on synthetic instances. More importantly, our experiments show that IDEQ performs very well on the instances of the TSPlib, a reference benchmark in the TSP community: it closely matches the performance of the best heuristics, LKH3, being even able to obtain better solutions than LKH3 on 2 instances of the TSPlib defined on 1577 and 3795 cities. IDEQ obtains 0.3% optimality gap on TSP instances made of 500 cities, and 0.5% on TSP instances with 1000 cities. This sets a new SOTA for neural based methods solving the TSP. Moreover, IDEQ exhibits a lower variance and better scales-up with the number of cities with regards to DIFUSCO and T2TCO.

For the early identification, diagnosis, and treatment of mental health illnesses, the integration of deep learning (DL) and machine learning (ML) have started playing a significant role. By evaluating complex data from imaging, genetics, and behavioral assessments, these technologies have the potential to improve clinical results significantly. However, they also present unique challenges relating to data integration and ethical issues. The development of ML and DL methods for the early diagnosis and treatment of mental health issues is reviewed in this survey. It examines a range of applications, with a particular emphasis on behavioral assessments, genetic and biomarker analysis, and medical imaging for the diagnosis of diseases like depression, bipolar disorder, and schizophrenia. Predictive modeling for illness development is further discussed in the review, focusing on the function of risk prediction models and longitudinal investigations. Important discoveries show how ML and DL might improve treatment outcomes and diagnostic accuracy while tackling methodological inconsistency, data integration, and ethical concerns. The study emphasizes the significance of building real-time monitoring systems for individualized treatment, improving data fusion techniques, and interdisciplinary collaboration. Upcoming studies should concentrate on surmounting these obstacles to maximize ML and DL's valuable and moral implementation in mental health services.

Control structure design is an important but tedious step in P&ID development. Generative artificial intelligence (AI) promises to reduce P&ID development time by supporting engineers. Previous research on generative AI in chemical process design mainly represented processes by sequences. However, graphs offer a promising alternative because of their permutation invariance. We propose the Graph-to-SFILES model, a generative AI method to predict control structures from flowsheet topologies. The Graph-to-SFILES model takes the flowsheet topology as a graph input and returns a control-extended flowsheet as a sequence in the SFILES 2.0 notation. We compare four different graph encoder architectures, one of them being a graph neural network (GNN) proposed in this work. The Graph-to-SFILES model achieves a top-5 accuracy of 73.2% when trained on 10,000 flowsheet topologies. In addition, the proposed GNN performs best among the encoder architectures. Compared to a purely sequence-based approach, the Graph-to-SFILES model improves the top-5 accuracy for a relatively small training dataset of 1,000 flowsheets from 0.9% to 28.4%. However, the sequence-based approach performs better on a large-scale dataset of 100,000 flowsheets. These results highlight the potential of graph-based AI models to accelerate P&ID development in small-data regimes but their effectiveness on industry relevant case studies still needs to be investigated.

The advance of large language models (LLMs) has unlocked great opportunities in complex multi-modal data management tasks, particularly in question answering (QA) over complicated multi-table relational data. Despite significant progress, systematically evaluating LLMs on multi-table QA remains a critical challenge due to the inherent complexity of analyzing the modality of relational data structures and the potentially large scale of serialized tabular data. Existing benchmarks primarily focus on single-table QA, failing to capture the intricacies of connections across multiple relational tables, as required in real-world domains such as finance, healthcare, and e-commerce. We present TQA-Bench, a long-context analytical multi-table QA benchmark derived from real-world public datasets, with a flexible sampling mechanism that varies context length (8K--64K tokens) and symbolic extensions for assessing reasoning beyond retrieval and pattern matching. We systematically evaluate a set of LLMs spanning model scales from 2 billion to 671 billion parameters. Our extensive experiments reveal critical insights into the performance of LLMs in multi-table QA, highlighting both challenges and opportunities for advancing their application in complex, data-driven environments.

Offline batch inference, which leverages the flexibility of request batching to achieve higher throughput and lower costs, is becoming more popular for latency-insensitive applications. Meanwhile, recent progress in model capability and modality makes requests more diverse in compute and memory demands, creating unique opportunities for throughput improvement by resource overlapping. However, a request schedule that maximizes resource overlapping can conflict with the schedule that maximizes prefix sharing, a widely-used performance optimization, causing sub-optimal inference throughput. We present BlendServe, a system that maximizes resource utilization of offline batch inference by combining the benefits of resource overlapping and prefix sharing using a resource-aware prefix tree. BlendServe exploits the relaxed latency requirements in offline batch inference to reorder and overlap requests with varied resource demands while ensuring high prefix sharing. We evaluate BlendServe on a variety of synthetic multi-modal workloads and show that it provides up to $1.44\times$ throughput boost compared to widely-used industry standards, vLLM and SGLang.

The generalization error curve of certain kernel regression method aims at determining the exact order of generalization error with various source condition, noise level and choice of the regularization parameter rather than the minimax rate. In this work, under mild assumptions, we rigorously provide a full characterization of the generalization error curves of the kernel gradient descent method (and a large class of analytic spectral algorithms) in kernel regression. Consequently, we could sharpen the near inconsistency of kernel interpolation and clarify the saturation effects of kernel regression algorithms with higher qualification, etc. Thanks to the neural tangent kernel theory, these results greatly improve our understanding of the generalization behavior of training the wide neural networks. A novel technical contribution, the analytic functional argument, might be of independent interest.

Deep learning models have shown strong performance in load forecasting, but they generally require large amounts of data for model training before being applied to new scenarios, which limits their effectiveness in data-scarce scenarios. Inspired by the great success of pre-trained language models (LLMs) in natural language processing, this paper proposes a zero and few shot load forecasting approach using an advanced LLM framework denoted as the Chronos model. By utilizing its extensive pre-trained knowledge, the Chronos model enables accurate load forecasting in data-scarce scenarios. Simulation results across five real-world datasets demonstrate that the Chronos model significantly outperforms nine popular baseline models for both deterministic and probabilistic load forecasting with various forecast horizons (e.g., 1 to 48 hours), even though the Chronos model is neither tailored nor fine-tuned to these specific load datasets. Notably, Chronos reduces root mean squared error (RMSE), continuous ranked probability score (CRPS), and quantile score (QS) by approximately 7.34%-84.30%, 19.63%-60.06%, and 22.83%-54.49%, respectively, compared to baseline models. These results highlight the superiority and flexibility of the Chronos model, positioning it as an effective solution in data-scarce scenarios.

Large Language Models (LLMs) hold great promise to revolutionize current clinical systems for their superior capacities on medical text processing tasks and medical licensing exams. Meanwhile, traditional ML models such as SVM and XGBoost have still been mainly adopted in clinical prediction tasks. An emerging question is: Can LLMs beat traditional ML models in clinical prediction? Thus, we build a new benchmark ClinicalBench to comprehensively study the clinical predictive modeling capacities of both general-purpose and medical LLMs, and compare them with traditional ML models. ClinicalBench embraces three common clinical prediction tasks, two databases, 14 general-purpose LLMs, 8 medical LLMs, and 11 traditional ML models. Through extensive empirical investigation, we discover that both general-purpose and medical LLMs, even with different model scales, diverse prompting or fine-tuning strategies, still cannot beat traditional ML models in clinical prediction yet, shedding light on their potential deficiency in clinical reasoning and decision-making. We call for caution when practitioners adopt LLMs in clinical applications. ClinicalBench can be utilized to bridge the gap between LLMs' development for healthcare and real-world clinical practice.

The task of music-driven dance generation involves creating coherent dance movements that correspond to the given music. While existing methods can produce physically plausible dances, they often struggle to generalize to out-of-set data. The challenge arises from three aspects: 1) the high diversity of dance movements and significant differences in the distribution of music modalities, which make it difficult to generate music-aligned dance movements. 2) the lack of a large-scale music-dance dataset, which hinders the generation of generalized dance movements from music. 3) The protracted nature of dance movements poses a challenge to the maintenance of a consistent dance style. In this work, we introduce the EnchantDance framework, a state-of-the-art method for dance generation. Due to the redundancy of the original dance sequence along the time axis, EnchantDance first constructs a strong dance latent space and then trains a dance diffusion model on the dance latent space. To address the data gap, we construct a large-scale music-dance dataset, ChoreoSpectrum3D Dataset, which includes four dance genres and has a total duration of 70.32 hours, making it the largest reported music-dance dataset to date. To enhance consistency between music genre and dance style, we pre-train a music genre prediction network using transfer learning and incorporate music genre as extra conditional information in the training of the dance diffusion model. Extensive experiments demonstrate that our proposed framework achieves state-of-the-art performance on dance quality, diversity, and consistency.

We propose a general framework for end-to-end learning of data structures. Our framework adapts to the underlying data distribution and provides fine-grained control over query and space complexity. Crucially, the data structure is learned from scratch, and does not require careful initialization or seeding with candidate data structures/algorithms. We first apply this framework to the problem of nearest neighbor search. In several settings, we are able to reverse-engineer the learned data structures and query algorithms. For 1D nearest neighbor search, the model discovers optimal distribution (in)dependent algorithms such as binary search and variants of interpolation search. In higher dimensions, the model learns solutions that resemble k-d trees in some regimes, while in others, they have elements of locality-sensitive hashing. The model can also learn useful representations of high-dimensional data and exploit them to design effective data structures. We also adapt our framework to the problem of estimating frequencies over a data stream, and believe it could also be a powerful discovery tool for new problems.

Generating lifelike human motions from descriptive texts has experienced remarkable research focus in the recent years, propelled by the emerging requirements of digital humans.Despite impressive advances, existing approaches are often constrained by limited control modalities, task specificity, and focus solely on body motion representations.In this paper, we present MotionGPT-2, a unified Large Motion-Language Model (LMLM) that addresses these limitations. MotionGPT-2 accommodates multiple motion-relevant tasks and supporting multimodal control conditions through pre-trained Large Language Models (LLMs). It quantizes multimodal inputs-such as text and single-frame poses-into discrete, LLM-interpretable tokens, seamlessly integrating them into the LLM's vocabulary. These tokens are then organized into unified prompts, guiding the LLM to generate motion outputs through a pretraining-then-finetuning paradigm. We also show that the proposed MotionGPT-2 is highly adaptable to the challenging 3D holistic motion generation task, enabled by the innovative motion discretization framework, Part-Aware VQVAE, which ensures fine-grained representations of body and hand movements. Extensive experiments and visualizations validate the effectiveness of our method, demonstrating the adaptability of MotionGPT-2 across motion generation, motion captioning, and generalized motion completion tasks.

It is becoming increasingly common in regression to train neural networks that model the entire distribution even if only the mean is required for prediction. This additional modeling often comes with performance gain and the reasons behind the improvement are not fully known. This paper investigates a recent approach to regression, the Histogram Loss, which involves learning the conditional distribution of the target variable by minimizing the cross-entropy between a target distribution and a flexible histogram prediction. We design theoretical and empirical analyses to determine why and when this performance gain appears, and how different components of the loss contribute to it. Our results suggest that the benefits of learning distributions in this setup come from improvements in optimization rather than modelling extra information. We then demonstrate the viability of the Histogram Loss in common deep learning applications without a need for costly hyperparameter tuning.

This paper proposes a non-lexicalized grammar-induction procedure that separates two tests: recognition of the observed finite presentation, and rejection of short preterminal strings generated by a hypothesis but unsupported by the evidence. The central object is the rule-coverage bound \(\ell^*(G)\): the maximum, over rules in \(G\), of the length of the shortest preterminal string whose derivation uses that rule. This bound induces the comparison universe \(Σ_{\mathrm{pre}}^{\le \ell^*(G)}\), where unsupported generated strings serve as indirect evidence against overgenerating hypotheses. We give a greedy search algorithm over rule sets and prove a conditional weak-recovery theorem: under explicit reachability conditions and sufficient saturation of the presentation, the exact learner reaches a grammar weakly equivalent to the unknown target. The complexity analysis is slice-wise: for each fixed incrementality radius \(k\), the search explores polynomially many rule-set extensions in the finite rule universe. Across 31 benchmark runs spanning Dyck-\(k\) languages \((1\le k\le4)\), palindromes, \(a^n b^n\), English-like recursive fragments, and an inherently ambiguous union language, grammar-level analysis establishes weak equivalence between every returned grammar and its target.

Past research relates design creativity to 'divergent thinking,' i.e., how well the concept space is explored during the early phase of design. Researchers have argued that generating several concepts would increase the chances of producing better design solutions. 'Variety' is one of the parameters by which one can quantify the breadth of a concept space explored by the designers. It is useful to assess variety at the conceptual design stage because, at this stage, designers have the freedom to explore different solution principles so as to satisfy a design problem with substantially novel concepts. This article elaborates on and critically examines the existing variety metrics from the engineering design literature, discussing their limitations. A new distance-based variety metric is proposed, along with a prescriptive framework to support the assessment process. The framework measures the real-valued distance between two design concepts using any chosen representation of their underlying abstraction levels. The proposed framework is implemented in a software tool called 'VariAnT.' Furthermore, the tool's application is demonstrated through an illustrative example.