arXiv subjects

Chao Ma

Publications and source records attributed to Chao Ma.

At least 19 recordsLinked to original sources

arXiv2026

Nonreciprocal coupling enables tunable exceptional-point response in non-Hermitian circuits

Exceptional points (EPs) exhibit characteristic square-root spectral responses to weak perturbations, with response magnitudes governed by configuration-dependent prefactors that are often relatively small. Modulating these prefactor while preserving the exceptional characteristics offers an additional degrees of freedom for engineering non-Hermitian sensing responses. Here, we report a nonreciprocal non-Hermitian circuit system with gyrator-meditated coupling that supports a continuous family of EP configurations. Unlike reciprocal coupling, which typically yields a unique EP configuration for balanced gain and loss, nonreciprocal coupling allows the directional forward and backward coupling strengths to be varied while maintaining their product and thus preserving the EP condition and yielding a library of EPs. This additional freedom enables continuous control of the frequency-splitting magnitude without modifying the square-root characteristic dependence on absolute perturbation strength. We show that the spectral splitting induced by a given absolute resistance perturbation can be enhanced by appropriately adjusting the coupling asymmetry. Theoretical predictions, circuit simulations, and experimental reflection-spectrum measurements consistently confirms this tunable response. Our results establish nonreciprocal coupling as an alternative approach for engineering and enhancing EP-induced perturbative responses without changing their characteristic branch-point scaling.

physics.optics↗

arXiv2026

Tensor products of Leibniz bimodules and Grothendieck rings -- an addendum

We continue our investigation of certain associativity properties (namely, alternativity, power-associativity, or being a Jordan ring) of the non-associative multiplication defined on the Grothendieck group of the category of finite-dimensional Leibniz bimodules which we started in \cite{FW}. Namely, we prove that for every finite-dimensional non-zero Leibniz algebra over an arbitrary field this Grothendieck ring is not power-associative. As a consequence, the Grothendieck ring is also neither alternative nor a Jordan ring. This corrects and generalizes some of our results in the last section of \cite{FW}.

math.RA↗

arXiv2026

MEgoVista: Multi-view Ego-aware Motion Estimation for Metric 4D Hands and Head in the Wild

Learning manipulation from human video requires high-fidelity hand-motion reconstruction in metric units. Today's metric hand labels come from studio rigs and instrumented headsets, and both are confined in the same two ways: neither leaves a prepared setting, and neither is checked against an independent reference. Unconstrained head-worn recording promises the opposite trade-off, scaling with the number of people wearing a device. We therefore introduce MEgoVista, an offline pipeline that turns a single unprepared MEgo View recording into metric two-hand and head motion in one gravity-aligned world frame. Three properties set it apart from existing egocentric reconstruction systems: first, it reconstructs in settings studio volumes and tabletop rigs cannot reach, settling hand ownership at detection so bystander hands stay out of the wearer's trajectory; second, it takes its metric gauge from calibrated stereo rather than a monocular prior, installing scale at initialisation so policies receive physical units, not arbitrary coordinates; third, both outputs are scored inside a motion-capture volume against independent Chingmu optical capture, under a protocol that audits its own reference and charges what a method declines to predict. MEgoVista is offered as a measured route from egocentric video to metric hand supervision, one that widens where such labels can be gathered. To bridge the gap between lab validation and industrial-scale data annotation, we further introduce MEgoDeploy, a production-grade GPU stack that fully preserves reconstruction fidelity while achieving a 14.05x end-to-end throughput speedup, scaling the generation of metric 4D pseudo-labels for embodied foundation models.

cs.CV↗

arXiv2026

BluePixel: Toward Native 4K Agentic Image Editing

Instruction-driven image editing has reached native 4K, yet how such editors are trained remains undisclosed, and pretraining one from scratch is beyond most teams. The task itself compounds the difficulty: a single end-to-end editor cannot make precise parametric adjustments, match specialized experts, or interpret the vague instructions users actually write. We argue that 4K editing is a system problem rather than a single-model problem, and present BluePixel, which couples a native-4K editing model with an agent harness; the complete training recipe is disclosed in this paper. The model makes 4K training practical: it turns a 1K-pretrained editor into a native-4K one through progressive VAE adaptation that keeps 4K training as affordable as 1K, absorbs a long tail of vertical editing skills through continual learning, and refines the aesthetic-fidelity trade-off with reinforcement learning. Distillation via source-anchored target projection (SATP) then compresses 50 denoising steps to 4, completing a 4K edit in 8.7 s on a single Kunlunxin P800 XPU while preserving task success rate and background consistency. The harness, BluePixel Agent, makes editing controllable: it decomposes user instructions into specialist and general operations, verifies every result with the same reward model that guided training, and retries within bounded attempts. On VINS-4KEval++ (3,305 native-4K cases), BluePixel ranks second among seven models, behind only GPT Image 2 in non-edited-region consistency.

cs.CV↗

arXiv2026

State-OPD: State-Grounded On-Policy Distillation for Multi-Turn Agents

On-Policy Distillation (OPD) has emerged as a powerful paradigm for transferring reasoning capabilities from frontier models to smaller models, while extending its success to multi-turn agentic environments remains challenging. We study agentic OPD through interaction states, distinguishing the objective task and executed action-observation record from the internal semantic trajectory of student reasoning. Through this lens, we identify two roles of student reasoning in multi-turn OPD: prior reasoning distorts teacher supervision, while current reasoning shapes the action alternatives sampled at the same state. Specifically, at matched interaction states, removing prior student reasoning improves teacher continuation success and changes the teacher's token-level targets, with larger continuation gains at later turns. Meanwhile, resampling complete reasoning-action responses yields greater action diversity at the same state than resampling actions under a fixed reasoning trace. We introduce State-Grounded On-Policy Distillation, a framework that grounds teacher supervision in interaction states and reallocates the sampling budget to generate multiple complete reasoning-action responses at states with high response entropy, while retaining full-history student generation. Comprehensive evaluations across ALFWorld, ScienceWorld and WebShop demonstrate that State-OPD achieves the highest macro-average success rate among the evaluated methods for both Qwen3-1.7B and Qwen3-4B students. Notably, for Qwen3-4B, State-OPD raises macro-average success from 26.07\% with the strongest baseline to 31.74\% and surpasses its Qwen3-32B teacher on ALFWorld Seen. Further analyses reveal that State-OPD achieves higher task success with fewer supervised turn responses and substantially reduces persistent failed-action repetition; qualitative traces illustrate decision revision based on interaction feedback.

cs.LG↗

arXiv2026

Extensions of Restricted Simple Modules for Restricted Lie Algebras of Cartan Type

We determine the first extension groups between simple restricted modules for the restricted simple Lie algebras of Cartan type over an algebraically closed field of characteristic $p\ge5$, which answers the restricted case of a problem of Benkart and Feldvoss. Extensions from parabolically induced modules are computed by finite restricted-cohomology complexes of the grading parabolic, and for the exceptional modules we obtain closed formulas in all four families. Neither the trivial module nor any nontrivial exceptional simple has self-extensions. In characteristic seven we determine the quiver of $S(3;1)^{(1)}$, which has 49 vertices and 566 arrows.

math.RA↗

arXiv2026

Augmentation-Ideal Realization of Special Nottingham Congruence Series and Cartier Indecomposables

The shifted congruence series of finite special Nottingham jet groups is realized by an integral augmentation filtration. For volume-preserving formal substitutions in characteristic $p\ge5$ and $n\ge3$ variables, the filtration induced by the action on truncated power series has the congruence subgroups as its dimension subgroups. The canonical weighted augmentation filtration has the same dimension subgroups. The group symbols form a degree-truncated Lie algebra of divergence-free polynomial fields. Since brackets with quadratic fields give the subspace of exact fields in every coefficient degree at least three, the Lie indecomposables occur only in degrees $(p-1)(n-1)+pr$, with $r\ge0$, where they are determinant-weighted Frobenius twists of tensor products of symmetric powers with the defining module.

math.RA↗

arXiv2026

Extremal Derived Length in the Unipotent Radicals of Cartan-Type Automorphism Groups

Over perfect fields of characteristic $p\ge5$, the unipotent radicals of the height-one special, Witt-Jacobson, and Hamiltonian automorphism groups in at least three, two, and four variables, respectively, attain the largest derived length allowed by their nilpotency classes. More strongly, every term of each derived series equals the corresponding power-of-two term of the lower central series. These equalities hold on finite-field points, with explicit Frattini quotients and lower $p$-central series. For the special family, a Cartier quotient measures the difference between the congruence and lower central filtrations. The Hamiltonian correction comes from flux and the socle line; the Witt-Jacobson family has none. We also identify the contact automorphism group scheme and, at arbitrary divided-power height, compute the Cartier-Verschiebung maps on restricted derivations and the Verschiebung filtration and distribution algebra of the Frobenius thickening of the special radical.

math.RA↗

arXiv2026

Lazard's Realization Problem for N-Series: Bar Obstructions and Finite-Quotient Descent

Lazard asked which $N$-series of a group arise from multiplicative filtrations of the integral group ring. Given an $N$-series $H_\bullet$ of a group $G$, let $A_\bullet$ be the filtration of the augmentation ideal spanned by products of elements $x-1$, $x\in H_a$, weighted by $a$. Every realizing filtration contains $A_\bullet$, so $H_\bullet$ is realizable exactly when the maps $H_n/H_{n+1}\to A_n/A_{n+1}$, $h\mapsto h-1$, are injective. We show that each kernel is the cokernel of an explicit map of normalized bar groups; for a finite group this makes realizability decidable by integer linear algebra. For a profinite group with a cofinal tower of finite quotients and an $N$-series induced from it, suppose $h\in H_n$ and $h-1\in A_{n+1}$ in every finite quotient. Then $h-1\in A_{n+1}$ in the discrete group ring exactly when the widths of its finite-level Losey expressions are bounded. This descent can fail. Using Tahara's class-three $2$-groups and the Hartl-Mikhailov-Passi description of the fourth dimension subgroup, we construct a countable product $P$ of finite $2$-groups. It contains an element of $γ_3(P)$ that lies in the fourth dimension subgroup of every finite subproduct of $P$ but not in $D_4(P)$.

math.GR↗

arXiv2026

The Derived Algebra of Nonlinear Polynomial Divergence-Free Vector Fields

We study the Lie algebra $L_{\ge2}$ of divergence-free polynomial vector fields on $k^n$, $n\ge3$, with coefficients of degree at least two, graded by coefficient degree. Over every field its derived algebra in degree $d\ge3$ is the space of exact fields, those whose contraction with the volume form is an exact form, and it is already spanned by brackets with quadratic fields. In characteristic zero this is the whole degree-$d$ component. In characteristic $p>0$ the abelianization is nonzero above degree two exactly in the degrees $d\ge(p-1)(n-1)$ with $d\equiv1-n\pmod p$, and Cartier descent identifies it with a Frobenius twist of a rational $\mathrm{GL}_n$-module, tensored with a power of the determinant. For $p\ge5$ each exact component is obtained from the previous one by bracketing with quadratic fields.

math.RA↗

arXiv2026

AtlasNav: Mitigating Evidence Blindness with Persistent Corpus Navigation

As language-model agents become more capable of iterative search, corpus access is shifting from retrieval toward interaction. Agents can explore the corpus, inspect documents, and use newly discovered evidence to decide what to examine next. Yet accessible evidence may still fail to become usable within a finite interaction budget. We call this progressive failure Evidence Blindness: supporting documents may never enter view, may remain unopened, or may fail to expose the decisive evidence even after being opened. A key reason is that agents often have to infer useful evidence directions during interaction, spending limited budget on deciding where to search next. Existing approaches either leave corpus structure largely implicit or reconstruct useful directions at query time. AtlasNav instead organizes reusable cross-document structure before any query arrives. It builds a persistent multi-view Corpus Atlas, which each query can navigate adaptively while still accessing the original documents directly. On BrowseComp-Plus, AtlasNav outperforms the previous state-of-the-art interactive corpus access method across different backbones. On DeepSeek, it improves strict accuracy by 7.47 points while reducing query-time inference cost by 30.22%.AtlasNav also reduces Evidence Blindness, realizes complete evidence earlier, remains robust to corpus-structure and scale shifts on PhantomWiki, and achieves leading performance on heterogeneous enterprise data.

cs.AI↗

arXiv2026

Large Knowledge Model: A Knowledge Foundation for Agentic Science at Scale

Agentic science envisions many autonomous agents investigating concurrently while building on a shared, evolving body of scientific knowledge. This requires a knowledge foundation that supports high-concurrency access, preserves traceable and reusable reasoning, and grows incrementally. We propose the Large Knowledge Model (LKM), a growing, agent-native knowledge foundation that provides a general representation of scientific knowledge across disciplines. LKM organizes the scientific literature into reasoning graphs, with claims as the core nodes and associated reasoning chains that make explicit how premises and evidence support conclusions. These source-grounded objects are persistent and addressable; cross-paper links organize them into aligned question, workflow, and evidence views. Newly extracted papers extend the foundation incrementally while preserving existing object identities. Building on this foundation, we develop an agent-native, reasoning-aware scientific retrieval system that retrieves claims together with their reasoning chains and sources, enabling agents to inspect and reuse the evidence underlying scientific conclusions. Across benchmarks, agents using LKM retrieve more evidence, cite more faithfully, and answer scientific questions more accurately: LKM nearly doubles the known supporting and contradicting evidence retrieved on SciFact-Open (818 versus 443 claim-paper pairs), reasoning graphs raise citation F1 on ScholarQABench by more than 5 points over the same retrieved papers, and LKM retrieval improves a fixed answering model by 9.3, 4.2, and 14.7 points over no retrieval on ChemBench, PubMedQA, and SciBench. LKM lays the foundation for a scientific ecosystem in which AI scientists not only recall accumulated knowledge but also extend it, returning new questions, workflows, and evidence to a memory that every subsequent investigation can build on.

cs.AI↗

arXiv2026

MOPD-Router: Rethinking Teacher Routing in Multi-Teacher On-Policy Distillation

Multi-teacher on-policy distillation (MOPD) integrates specialized capabilities into a single student, but existing practice typically hard-routes each prompt to a domain-matched teacher for the entire rollout. This dependence on prompt-level domain labels restricts using unlabeled training mixtures and leaves complementary signals from other teachers unused. We introduce MOPD-Router, a framework that routes supervision over the full teacher pool at each token, without domain labels or training a separate routing model. Its plug-in interface supports different metrics for selecting and weighting teacher-specific OPD signals. Within this interface, we propose ExpertAlign, which scores each teacher by whether its correction to the student at the current token expresses the specialization that teacher acquired during post-training, and compare it against two reference metrics built on teacher confidence (Entropy) and teacher-student discrepancy (Novelty). Experiments on unlabeled and domain-labeled training mixtures under strong-to-weak and same-size distillation scenarios show that ExpertAlign achieves the strongest overall performance in all four settings. On unlabeled data, it improves the overall score by 5.88 (+12.3%) points over Mean aggregation; on domain-labeled data, it outperforms standard MOPD by 3.95 (+7.8%) points without using available domain labels. These results demonstrate token-level routing can exploit cross-domain complementary supervision, and reduce exclusive reliance on prompt-level domain assignment. Code is available at: https://github.com/TURLEing/MOPD-Router.

cs.LG↗

arXiv2026

Counterfactual Constraint-Conditioned On-Policy Distillation for Multi-Constraint Instruction Following

Multi-constraint instruction following requires a model to respond to a query under many simultaneously active constraints. Even strong instruction-tuned models still routinely violate some of them. Existing approaches either augment supervision with sequence- or token-level RL rewards from external verifiers or learned graders, or use on-policy distillation (OPD) against a single full-context teacher whose probability mass becomes diluted as more constraints become simultaneously active. We propose CC-OPD (Counterfactual Constraint-Conditioned On-Policy Distillation), which inverts the standard supervision-generation direction in distillation. Rather than enriching the teacher with information beyond what the student sees, CC-OPD ablates each constraint from the teacher's conditioning in turn, and constructs the per-constraint signal from the resulting per-token probability differentials. The resulting per-token leave-one-out log-likelihood shifts are summed, clipped, and added to the vanilla OPD reward as a token-level shaping term. All shaping terms are obtained from the frozen teacher, without an external verifier during distillation, and the reward equals vanilla OPD wherever the aggregate shift is zero. Across two Qwen model pairs and seven benchmarks, CC-OPD achieves the highest average among all evaluated student-training methods. A 1.5B student trained with CC-OPD surpasses its own 7B RL-trained teacher on the MulDimIF benchmark.

cs.LG↗

arXiv2026

PLAT: Sparse Timed Keyframe Motion Tracking for Humanoid Control via Privileged Latent Transition Learning

Humanoid motion tracking policies rely on dense frame-by-frame references, limiting their use as high-level motion controllers for planning and interactive motion generation. We study \emph{Sparse Timed Keyframe Motion Tracking}, where a policy receives only sparse future keyframes and their desired arrival times, and must execute stable whole-body motions that reach successive goals. We propose \textbf{PLAT}, a three-stage sparse timed keyframe motion tracking policy learning framework with \textbf{P}rivileged \textbf{LA}tent \textbf{T}ransition learning. PLAT bridges dense motion tracking and sparse goal-conditioned control by exploiting dense goal sequences as privileged supervision during training while requiring only sparse timed keyframe commands at deployment. A pretrained dense tracking expert first provides robust motion priors. A privileged latent prior is then learned through DAgger-style imitation, followed by latent residual reinforcement learning that refines latent transitions instead of directly optimizing actions. Extensive simulation experiments demonstrate that PLAT maintains accurate and stable sparse timed keyframe tracking across varying planning horizons, with particularly strong performance under long-horizon commands. Successful deployment on a Unitree G1 humanoid robot further demonstrates the effectiveness and practicality of PLAT for sparse humanoid motion control.

cs.RO↗

arXiv2026

PointLAM: Local Attentive Mamba for Efficient Point-based 3D Object Detection

3D object detection from LiDAR point clouds faces a fundamental dilemma: voxel-based methods achieve efficiency at the cost of geometric quantization, while point-based methods preserve fidelity but suffer from prohibitive computational bottlenecks. Specifically, point-based architectures are crippled by slow downsampling strategies (e.g., FPS) and expensive dynamic neighbor queries (e.g., k-NN) coupled with costly continuous interactions. To tackle these systemic inefficiencies, we propose PointLAM, a highly efficient and powerful point-based architecture driven by two synergistic innovations. First, to resolve the downsampling bottleneck, we develop the Laplacian Point Sampler (LPS). LPS employs an implicit discrete Laplacian high-pass filter and Doubly Sorted Sampling to achieve fast, structure-aware foreground preservation. Second, to overcome local modeling latency, we design the Local Hadamard Aggregator (LHA). LHA decouples spatial indexing from feature representation using transient grids, and replaces complex continuous interactions with a Hadamard Gating mechanism for topology-aware, attentive modulation. By coupling this local gating with Bi-Directional Mamba (BDM) layers for global sequence modeling, we formulate the Local Attentive Mamba (LAM) block. Powered by this architecture, PointLAM achieves competitive performance on nuScenes and Waymo for point-based detectors. It rivals highly optimized voxel competitors while requiring a fraction of the computational footprint, demonstrating marked superiority in detecting small instances and handling extreme sparsity. Project page: https://pointlam.github.io/.

cs.CV↗

arXiv2026

Pushing HST to the limit: Detection completeness and morphology robustness of faint galaxies in the EGS field

Reliable catalogs at the faint limit are essential for statistical studies of galaxy evolution. However, detection and structural measurements of faint galaxies remain challenging. We evaluate detection completeness and structural parameter uncertainties of faint galaxies in the CANDELS/EGS field using mock images and observations from the HST and JWST. The framework is also applicable to large high-redshift surveys, including Euclid, the upcoming CSST, and the Nancy Grace Roman Space Telescope. We detect sources in the EGS F160W mosaic using a 1.5σthreshold and quantify detection completeness and false-detection rates with realistic mock images. Mock galaxy-pair tests assess blending effects on both source detection and structural measurements in sersic fitting. In addition, the overlap with CEERS observations is used to derive an empirical criterion for separating true faint galaxies from fake detections using HST photometry. We identify 72,467 sources in the CANDELS/EGS field, including 57,444 fainter than 25 mag in F160W. Detection completeness exceeds 95% for objects brighter than 25 mag but declines to <50% beyond 27 mag (with false detection rate>50%). Structure parameters from single sersic modeling are reliable down to mH = 25 mag. For galaxies at 25<mH<27 mag, systematic biases in the sersic index reach Δn \approx 5-70%, while Δq decreases from approximately -0.03 to -0.27. Close companions (~0.3 arcsec) introduce measurable biases in recovered sersic magnitudes and shapes, while the impact becomes negligible at larger separations. An empirical criterion based on mH, F814W-F160W color, R50, and zphot separates true faint galaxies from false detections, reaching 72% completeness and 84% purity. It yields a final catalog of 59,646 objects consistent with previous CANDELS measurements at bright magnitudes while reaching fainter limits.

astro-ph.GA↗

arXiv2026

The 2024 MRSI Data Processing and Quantification Challenge Synthetic Dataset

Synthetic data is central to magnetic resonance spectroscopy method development because they provide ground truths for software validation, reproducible benchmarking, and machine- and deep-learning training. In MRSI, synthetic data must capture spatially varying anatomy, field inhomogeneity, nuisance signals, and measurement effects. The 2024 MRSI Data Processing and Quantification Challenge Synthetic Dataset is a simulated 3T brain FID-MRSI resource with ground-truth metabolite maps developed as a controlled testbed for MRSI processing and quantification methods. Subject-specific simulations used anatomical images and field maps from Human Connectome Project subjects. Tissue masks, quantum-mechanically simulated metabolite basis functions, in vivo-derived macromolecular components, Bloch-simulated post-WET residual water, registered in vivo lipid signals, spectral baseline, $B_0$-dependent frequency shifts, Voigt lineshape variations, and complex Gaussian noise were combined in a forward model. Water and lipid signals were synthesized on a high-resolution grid and Fourier-truncated to the final echo-planar spectroscopic imaging grid to model the finite spatial point spread function. This resource has 24 training and 8 testing datasets containing contaminated FID-MRSI data, anatomical images, $B_0$ maps, metadata, and ground-truth metabolite and component signals. Forward-model parameters are documented, including tissue-specific metabolite concentrations and relaxation times, macromolecular amplitude ratios, the water model, noise, and field-inhomogeneity ranges. This synthetic benchmark supports development and comparison of MRSI processing, nuisance-signal removal, and metabolite-quantification methods, while enabling method evaluation. It illustrates how characterized synthetic data can support rigorous evaluation when ground truth is difficult or impossible to obtain.

physics.med-ph↗