Science & Mathematics

Currently, dexterous humanoid robot hands are widely considered the final hardware bottleneck that hampers generalist robots from physically assisting people at home and work. As conventional tendon-driven designs suffer from high cost, demanding maintenance, poor reliability, and short durability, alternative linkage-driven designs have increasingly gained attention. Despite their potential to avoid the weaknesses of tendon-driven hands, integrating multiple degrees of actuation (DOA) in a single digit with linkage drivetrains is a challenge for design and fabrication. Only a few studies have managed to present functional humanoid hands that are completely linkage-driven and have at least three DOAs per digit without compromising actuator placement and other design preferences. Our work introduces a new linkage-driven hand featuring a unique flexure hinge. Together with elastic drive rods made of spring steel, the linkage drivetrains are significantly simplified. Unlike other linkage-driven designs of similar dexterity, our design does not require many miniature moving parts. As a result, the hand can be fabricated easily with a budget 3D printer. This not only significantly lowers the barrier to entry for using and developing dexterous humanoid robot hand functions, but also allows the hands to be sized much closer to regular human hands. The design offers desirable kinematic advantages as well. A prototype is evaluated for grasping and object manipulation. In addition, this paper presents our latest findings on kinematics and control of the hand.

Kunqu is one of the oldest extant forms of Chinese opera, yet its digitization remains scarce and difficult to annotate, with two challenges standing out: first, close analysis of Kunqu performance demands syllable-level lyrics alignment; second, analyzing Kunqu pronunciation requires its phonological attributes (e.g., rhyme) to be linked with each syllable. To address these challenges, we present KunSet, the first Kunqu performance corpus comprising 482,164 fine-grained annotations of syllable-level lyrics and phonological attributes in total media duration of 155.2 h with 106.3 h annotated. To ensure annotation quality, KunSet was preannotated by an iterative ML-assisted pipeline and corrected by over 200 trained annotators via a purpose-built online annotation system under a two-round review process. To assess KunSet as a music information retrieval (MIR) benchmark, we evaluate two tasks: lyrics alignment, by finetuning a forced-alignment model with an acoustic anchor decoder; and same-rhyme retrieval, using frozen Chinese HuBERT features. Models built with KunSet outperform their baselines. We release KunSet and model checkpoints to support MIR research on traditional-music corpora.

In this work we formalize a new class of binary trees called encoding trees. We extend this structure in two ways, through labelling to create the class of labelled encoding trees, and pruning to create the class of pruned encoding trees. We discuss how these two extensions can be combined to create the the class of pruned labelled encoding trees. Using the supplementary results that we prove about these classes of trees, we construct a code that uses pruned labelled encoding trees to convert a string of symbols from a sequence of non-empty sets to a binary string. We prove that this code is a variable-length prefix code that can process any valid string of symbols from the given sequence of non-empty sets. Furthermore, we construct a code from the class of pruned labelled encoding trees that goes in the opposite direction, from binary strings to a string of symbols from a sequence of non-empty sets of arbitrary symbols. We prove that this code is also a variable-length prefix code that can process any binary string of any length, provided that the sequence of non-empty sets is sufficiently long. We provide an equation that establishes the minimum sequence length to encode any binary string of a given length, based on that length.

Dropout and zoneout are routinely added to long short-term memory (LSTM) and gated recurrent unit (GRU) networks, but, to our knowledge, their placements have not been crossed with gate initialisation in a factorial design, and it is not known whether early gradient propagation tells which runs will learn long dependencies. We trained one-layer LSTMs and GRUs on the adding problem and on delayed recall in a factorial grid of 720 runs, crossing conventional and chrono gate initialisation with up to six recurrent regulariser placements at one rate each, with pre-specified follow-ups, and recorded how sensitive the final hidden state was to the input at every lag. Because both tasks supply fresh or exhaustive training data, the study measures the regularisers' effects on optimisation and does not assess generalisation from a finite training set. At lengths chosen to make conventional initialisation difficult, all 360 chrono-initialised runs reached the validation criterion within 5000 updates, as did 119 of 120 unregularised runs with a horizon about four times too short or too long, whereas conventionally initialised networks learned slowly or did not reach the criterion within the budget at the longer lengths. Under conventional initialisation, dropout generally delayed learning and zoneout helped only in some conditions. The early input sensitivity added little to predicting which conditions learn once the early validation loss was modelled per task, architecture and initialisation. Within conditions, a flatter early decay went with the seeds that learned, but in a pre-specified test the association was modest (concordance 0.61), condition-dependent and not robust to resampling seeds. The results support chrono initialisation, even with a rough horizon, on such tasks.

Explaining a time-series model requires both locating the evidence and extracting meaningful knowledge from it. Yet a single timestep is not a semantic unit: the meaningful patterns live across many samples, not in any one of them. Per-timestep saliency is therefore noisy, high-dimensional, and unstable. Existing methods address this by introducing fixed-size windows and treating them as independent players in a cooperative game. Evaluating a coalition of such players requires defining a ``baseline'' region, which prior work tackles with fixed counterfactuals (e.g., a flat line, noise, or a spectral surrogate). This choice frequently dominates the resulting explanation, obscuring the model's true decisions. To address these issues, we propose GRAIL, a region-based time-series explanation framework that uses domain-specific intervals, such as the P-wave and the QRS complex of an electrocardiogram (ECG) beat, as players. We describe each region with human-readable features and generate counterfactual baselines with an on-manifold conditional generator. Such a choice of players also admits non-temporal features: static covariates enter as singleton players attributed with the temporal regions on one scale, so a patient's tabular evidence (e.g., age, sex) is weighed directly against the signal. We attribute the predictions to the features via Owen values with common-random-number variance reduction. Experimental results on both synthetic and real data show that the proposed framework produces faithful and robust explanations, evaluated under matched on- and off-manifold operators, offering meaningful knowledge to domain experts.

We study the problem of characterizing the optimal consistency--robustness tradeoff for single-machine non-clairvoyant scheduling with predicted processing times. For a fixed number of jobs $n$, Wei and Zhang (2020) established a lower bound on the achievable consistency--robustness tradeoff which is known to be tight only in a few restrictive cases, leaving the optimal tradeoff for general $n$ unresolved. A different worst-case-$n$ regime was introduced by Benomar et al. (2025), who applied the former work's construction to show that for $λ\in (0, 1)$, any $(1+λ)$-consistent algorithm must be $1 + Ω(\sqrt{1/λ})$-robust. On the other hand, the best known upper bound gives $(1+λ)$-consistency and $1+O(1/λ)$-robustness, leaving a polynomial gap. In this work, we develop a new lower bound using a multi-stage adversarial construction and a new algorithm that together give a near-optimal characterization of the consistency--robustness tradeoff: for every $n$, our upper bound is within a multiplicative factor $3/2$ of our lower bound, and thus our algorithm is within a multiplicative factor $3/2$ of the online optimal. In particular, our lower bound shows that any $(1+λ)$-consistent algorithm must have robustness at least $1/\sqrtλ$ as $n \to \infty$, while our algorithm achieves robustness at most $\frac{1}{2}+\sqrt{\frac{5}{4}+1/λ}$ for any $n$, matching the lower bound in the leading term as $λ\downarrow 0$; our upper bound also closes the polynomial gap in the worst-case-$n$ regime. We additionally show how our algorithm can be made error-tolerant, preserving the near-optimal consistency-robustness tradeoff while incurring only slightly degraded performance under bounded prediction errors.

We introduce HumanCentric-Scenes-AI, a human-curated dataset of 296 AI-generated stimuli that depict people across four contexts in which people encounter AI-generated media. The Surveillance subset contains 85 AI-generated CCTV-style images. The Occupations subset contains 128 AI-generated occupational portraits in a balanced paired design covering 64 occupations with one female and one male version of each. The AI-edited subset contains 27 real photographs in which a region was altered or replaced by a generative model. The Video subset contains 56 AI-generated short videos. We curated the images and videos manually over multiple rounds to make them as photorealistic as possible. For each stimulus we release the file, the generation prompt where recorded, and per-stimulus metadata. We report depicted-gender distributions from an initial set of 134 generations without a specified gender and compare them with U.S. Bureau of Labor Statistics occupational gender shares. We report edit-region size and its association with zero-shot detection difficulty for the AI-edited subset. We report image-quality distributions for the Surveillance subset and subject-category coverage for the Video subset. HumanCentric-Scenes-AI is intended first for human-perception research, and also for calibration of AI-detection classifiers and bias analysis of generative models.

Large language models (LLMs) have shown considerable promise in synthesizing formal specifications from source code. However, existing approaches primarily focus on functional contracts and loop invariants, while largely neglecting frame conditions, which specify the memory locations that may or may not be modified during execution and constitute an essential part of formal specifications. To address this limitation, we propose an LLM-based approach for synthesizing ACSL frame specifications alongside functional contracts from C code. Our approach analyzes code property graphs (CPGs) to extract frame-relevant information from the input code and leverages this information to guide the LLM in generating accurate frame conditions. The resulting specifications capture both region-separation constraints and data dependencies that govern variable modifications. Experimental results demonstrate the effectiveness of our approach. On a benchmark of 84 ACSL-compliant functions from non-trivial real-world systems, our method successfully synthesizes error-free frame specifications for 92.9% of the functions, representing a 179% relative improvement over state-of-the-art techniques. On a second benchmark consisting of 51 functions, the success rate increases from 80.4% to 96.1%. Furthermore, when integrated as an auxiliary component into existing LLM-based ACSL-synthesis frameworks, our approach improves frame-specification generation performance by up to 61.9%.

Understanding the properties of fish populations is increasingly important. Analyzing fish length, width, and mass is integral to maintaining healthy stock in the aquaculture industry, while counting wild fish populations supports marine conservation and coastal reef restoration efforts. Deep learning-based segmentation models offer a promising means of obtaining such information at scale. Most of these models, however, require costly and time-consuming per-pixel segmentation labels for training, which can take up to two minutes per fish to annotate manually. To address this inefficiency, we propose a weakly supervised segmentation pipeline that requires only point-labeled data (one click per fish) to produce accurate segmentation masks. Our pipeline consists of four stages. First, a point-supervised tracker predicts point annotations for each frame. These points are then converted into motion-informed language expressions, which guide a language-conditioned segmentation model (LMPM++) to capture temporal consistency and produce candidate segmentation masks. Finally, optical flow and background subtraction are applied as post-processing filters to suppress noise from incorrect segmentations. Experiments on the DeepFish dataset show improvement in foreground mean intersection over union (mIoU) as compared to baseline models.

Specular chains concentrate important transport in small regions of path space and the structure depends on chain topology. We introduce ART-MPG, a compact neural model that samples chain length, an autoregressive reflection/transmission (R/T) string, and a direction from a topology-conditioned four-component von Mises-Fisher (vMF) mixture. We extend the repeated trial weighting in Manifold Path Guiding (MPG) by conditioning retry sampling on both chain length and the sampled R/T string, accounting for its known mass outside the stochastic re-hit trials. For a fixed proposal and target chain, this preserves the ideal expectation while reducing the expected retry count and reciprocal-estimator variance compared with retries conditioned on chain length alone. We distinguish the formal unbiasedness result from the capped and clamped renderer used in experiments. We evaluate fourteen scenes across a range of samples per pixel (SPP) against existing online and offline rendering algorithms. Our method consistently produces lower mean squared error at equal SPP than the online and offline baselines, with lower standard deviation across runs.

A typed decision maps evidence, a question and a fixed set of options to a probability for every option, computed in one forward pass, by reading a language model's logits or a small head over its hidden states, instead of generating and parsing text. The commercial system Jev popularised this System~1 interface, and open reproductions have followed. We report two uses of the pattern. \emph{Part I} describes an in-house System~1 judge that triages about 148K agentic coding session records for a training corpus. It replaces a frontier LLM judge that cost on the order of \$1,000 per pass and whose stored labels matched a fresh two-judge consensus on only 26.9\% of the sessions where the two judges agreed; a self-hosted frozen backbone with a rank-16 LoRA adapter and a pointer head answers three typed questions in one forward pass, judges 1,000 sessions in about three minutes of one GPU, about \$0.06 at a reference rate and roughly 1\% of the frontier judge's cost, and keeps every session inside the organisation. \emph{Part II} asks whether a readout that adds no parameters can be competitive and whether training one helps. Applying the public Eikos recipe and corpus of about 25K decisions to three further backbones (1B, 2B and 35B), we read the frozen output layer of a LoRA-adapted decoder. On the 111 public hard items of JevBench our single-seed 35B, developed with those items in view, scores 79.3\%, not significantly different from the commercial reference (73.0\%, unadjusted $p=0.12$) and ahead of five open systems after Holm--Bonferroni correction; our 2B is ahead of all four kev checkpoints on the leaderboard. A single-seed ablation finds no detectable benefit from training the 26 letter rows of the output layer, and a retrain on a later corpus release moves the 2B's hard-tier accuracy by 11.7 points, which shows how little small evaluations resolve.

Testing quantum software is challenging because quantum states cannot be directly inspected. Variational quantum algorithms (VQAs) are a particularly important target as they are among the most promising approaches for near-term noisy quantum hardware. Many testing methods rely on classically computed results, which conflicts with the regime where quantum advantage is expected. Although prior studies mitigate this oracle problem, the core difficulty remains that existing approaches do not verify whether the execution deviates from the intended problem or whether the obtained solution is physically plausible. We propose a symmetry-oriented testing framework for VQAs that addresses this gap by using operators commuting with the target Hamiltonian as test specifications. The framework provides two complementary diagnostics: a commutator-based health check that monitors whether each execution step during optimization remains consistent with the symmetry of the target Hamiltonian, and a variance-based acceptance check that assesses the converged solution by examining whether the state is an eigenstate of the symmetry operators. Both diagnostics derive from the problem structure and do not require the exact solution. We evaluate the framework on representative user- and system-side faults, across multiple target systems, and under injected noise. Results show that the two diagnostics detect faulty behavior, provide informative signals of solution quality, and remain useful under noisy execution.

A learner's present predictions do not fully determine how it will learn next. We study the dynamics of hidden learning structure through two related operations. Fiber Regulation Theory (FRT) describes protected repair of existing function and the learning responses left available after repair. Fiber Acquisition Theory (FAT) describes the formation and renewal of trainable structure, distinguishing neutral installation, access to a learning signal, and additional functional directions. The two operations also constrain one another. A repair obligation can force a change in a learning channel, while acquired features can make new protected repairs feasible or reduce the minimum coefficient cost of existing ones. We characterize joint admissibility and response reachability, and give a finite supervised-learning construction in which the two operations together produce a favorable response absent from either operation alone. We implement protected head repair and fixed-capacity LoRA renewal in Qwen2.5-32B with AdamW, under replay and output distillation. Across two training trajectories differing in seed and within-task example order, the coordinated policy achieves higher teacher-forced Exact than either separate policy in all four comparisons. It also has the lowest macro NLL under both learning rules on the second trajectory. The implementations realize structure-conditioned self-regulation within specified information and action constraints: accessible state readouts and arrived evidence guide repairs and structural renewals, and ordinary learning continues from the modified state.

Users increasingly rely on large language models (LLMs) to analyse collections of independent inputs, such as product reviews or social media posts, and return a single aggregated answer. This setting requires multi-instance processing (MIP), where a model combines information from multiple instances into one prediction. We study a task in which the model must infer a binary label for each instance, and identify the most frequent label among the full instance set. The correct answer should be invariant to instance order. By evaluating fifteen LLMs across five datasets and seven instance order patterns, with each instance set held fixed across patterns, we find substantial order sensitivity. Accuracy can differ by about 90 percentage points between the best and worst order patterns, with models generally more accurate when majority-label instances appear at the start than when they appear in the middle or at the end. For subsets of models, available reasoning traces and gradient-based attribution reveal uneven emphasis across positions. Traces disproportionately mention early instances, while attribution scores are generally higher near the beginning and end than in the middle. These results show that instance order can systematically bias aggregated predictions, highlighting the need to evaluate MIP across different input arrangements.

Dense-grid spectrum search is computationally costly for multiple signal classification (MUSIC) direction-of-arrival (DOA) estimation, and low-precision evaluation can introduce numerical errors that distort candidate selection. This paper proposes precision-aware hierarchical MUSIC (PAH-MUSIC), a multistage framework that coordinates arithmetic precision, angular grids, and candidate windows in spectrum evaluation, while covariance estimation and eigendecomposition remain in FP64. Stage-dependent precision is assigned from coarse candidate discovery to final DOA estimation. Under bounded perturbations and local spectral growth, a conditional displacement bound gives a sufficient window-width condition for retained, correctly associated candidates. For a two-stage continuous cost model with fixed precision weights and window coefficients, a closed-form optimal coarse step and a cost-gain condition are derived. Simulations with two separated sources and 500 frames at each tested signal-to-noise ratio (SNR) from -10 to 15 dB show that FP8, FP16, and FP32 coarse searches followed by FP64 refinement achieve DOA root-mean-square errors close to a baseline using FP64 at both search stages. Relative to that baseline, the selected configurations reduce weighted spectrum-evaluation cost by 28.36\%--44.66\% under the adopted precision weights.