arXiv subjects

Peng Xie

Publications and source records attributed to Peng Xie.

At least 19 recordsLinked to original sources

arXiv2026

Where Should Language Sit in a Multimodal Model? Lessons from What Language Does to Human Perception and Cognition

Language models compute over tokens: language is their input, their output, and increasingly their internal representation. Whether language should keep all of these positions depends on what language does to the system that uses it. The one system with a century of data on that question is the human. We review what language does to human perception, the brain, and thought, and read the same evidence against multimodal models and language models. Throughout, we treat language as a compressor that runs on a shared codebook: a word is an index, the content is in the receiver, and a community maintains the codebook. In humans the compression is measurable, learning the codebook reorganizes the senses, and thought survives the loss of language. We then measure the rule that models apply when two cues disagree, with cue-conflict experiments on six vision-language models and two robot policies. Surviving cues are weighted in the order their reliabilities prescribe, at 11 to 82\% of the ideal observer's slope, and many answers copy the text. One policy family drops a cue that adds no information beyond the others rather than down-weighting it, another keeps it at a weight that fails when the cues conflict, and a visual cue that identifies the task in every training frame is never learned, because the language pathway already fits the data. Language models are the best current models of the human language network, and they have entered the human speech community, shifting word frequencies while alignment narrows their conceptual diversity. We close with seven implications for token-based systems. Language belongs at a model's boundary and in the shared codebook, as in the brain, not as its internal representation; the price of leaving the codebook inside is auditability.

cs.CL↗

arXiv2026

Sampling-based Certified Planning with Graphs of Convex Sets

Planners on graphs of convex sets return trajectories that are collision-free by construction, provided the convex regions are collision-free. The region generator only promises that property probabilistically, and no planner in the family verifies it. We report the first measurement of what the gap costs. On a scaled 14-DOF bimanual library, $3.2\%$ of interface samples are in collision, and a search-based GCS planner (\gcsstar) turns that volume error into a $62\%$ answer error: $18$ of $29$ pick-and-place queries return trajectories that drive the arms through the shelves, up to $91$\,mm deep, reported as successes. Repairing the library does not work; a ten times stricter acceptance contract, sums-of-squares certified regions, and uniform margins each destroy the connectivity planning needs before they deliver soundness. We instead build a planner that certifies its answers. It samples the overlaps and shared faces of the decomposition, prunes with an admissible informed bound, and verifies the one candidate each search round proposes, continuously, by a chain of clearance certificate balls with no resolution parameter; failures are repaired with local in-region detours, and the convex polish is re-verified. Head-to-head on all $29$ task queries it delivers zero invalid answers against $21$ for the reference, reaches its first certified answer in $0.11$\,s against $1.59$\,s for the reference's unverified one, and reproduces the reference optimum exactly on every query whose reference answer is physically valid.

cs.RO↗

arXiv2026

CompCPZ: Preserving Multi-Modal Intent in Language-Guided Robot Manipulation

A robot asked to "place the cup near the red plate or the blue plate" may reach the centroid between them and appear geometrically successful, while satisfying neither disjunct of the instruction. This silent semantic failure exposes a structural limitation of language-conditioned robot policies: representations that collapse a disjunctive instruction into a single connected set cannot preserve all feasible modes, and planners that commit to one action degrade under run-time mode uncertainty. We address this limitation with CompCPZ, a sound algebraic layer that language-conditioned learning systems wrap to recover multi-modal disjunctive representation, recursively composing per-primitive constrained polynomial zonotope enclosures along the language parse tree with distribution-free conformal coverage and sub-millisecond runtime. On a closed-loop ManiSkill3 tabletop-manipulation benchmark, CompCPZ outperforms convex set baselines, multi-peak decoders, and a zero-shot vision-language-action model (1,900/1,918 paired wins, p << 10^(-30)); the same compiler also transfers without retuning to planar real-robot trials on a Unitree Go2 quadruped under motion capture. These results suggest that compositional language grounding should be evaluated not only by reaching a decoded target, but by whether the represented feasibility set preserves the connected-component structure of the user's intent.

cs.RO↗

arXiv2026

Certifying Plans under Model Mismatch: A Trilemma for Reachability from Scarce Data

Sim-to-real policies are designed under nominal dynamics, but target-system trials may yield only a few isolated one-step transitions. We study pre-execution certification of a fixed control sequence, such as an action chunk produced by a learned policy. If the sequence reaches an unobserved state-input region, the observations remain consistent with target systems whose trajectories separate along it by an arbitrarily large amount. Any deterministic certifier sound for all of them must then decline to certify or return a reachable tube with arbitrarily large projected width. For bounded smooth classes of the target-nominal model error, we derive a finite plan-dependent projected-width lower bound. These results expose a trilemma among uniform trajectory containment, finite projected width, and unrestricted model-error behavior beyond the observations. ForeReach requires a supplied componentwise Lipschitz bound on the model error. Observed transition pairs can refute this declaration but cannot establish it outside the observed locations. Conditional on a valid declaration, our method constructs a set-membership envelope for the model error, propagates a zonotopic reachable tube, and certifies only when propagation remains within the certification domain and every projected tube slice avoids the unsafe set. In two benchmark systems, calibration baselines may remain narrow after losing trajectory containment outside data support, whereas our method declines to certify unsupported sequences and recovers certification when relevant target data and sufficient obstacle clearance are available.

cs.RO↗

arXiv2026

Probabilistic Reachable-Action Verification of Visuomotor Policies via Set-Based Training

Reachability analysis for visuomotor policies is difficult because large visual encoders make end-to-end set propagation computationally expensive and excessively conservative. We therefore freeze the visual encoder and confine set propagation to a low-dimensional interface between it and the downstream policy, with the interface set calibrated from held-out camera-pose perturbations. Propagating this set through the policy with zonotopes yields a terminal output-enclosure width that set-based training optimizes directly. During evaluation, camera-pose perturbations are sampled from the prescribed distribution, and rollout-level split conformal calibration converts the resulting action-deviation scores into a probabilistic reachable-action radius with finite-sample coverage. In controlled manipulation experiments, set-based training reduces this radius while preserving closed-loop task capability, and matched behavior-only, observational-consistency, and pointwise-adversarial controls all leave a larger radius.

cs.RO↗

arXiv2026

The Intruder Threshold: A Spectral Law for LoRA Fine-Tuning

LoRA fine-tuning can create intruder dimensions: new leading singular vectors of the updated weight matrix $W+BA$ that are nearly orthogonal to all pretrained singular vectors and that drive catastrophic forgetting. Since their discovery, no theory has predicted, layer by layer on measured spectra, when they appear. We derive a per-layer critical update strength $s^\ast=\bar\theta/(\gamma\sigma_1(BA))$, computed from the measured spectrum of $W$ alone through the rectangular spiked-deformation transform, together with an exact secular-equation characterization of the updated spectrum, with no fitted parameters. In a pre-specified study spanning four dense Transformer families, a state-space model, a mixture-of-experts model, and an encoder-decoder (18 adapters, 9{,}840 layer scans), the law localizes the empirical threshold within a factor of two on $82\%$ of layers, separates intruder-bearing from intruder-free layers at deployment with a mean AUC of $0.89$, holds unchanged on six third-party adapters, and predicts where WikiText-2 perplexity begins to degrade; a combination of the two pre-specified edge evaluations reaches $98\%$ and is confirmed out-of-bag on the external adapters ($0.997$). Full fine-tuning disperses its update far below the threshold of every layer, which resolves the asymmetry between LoRA and full fine-tuning. Norm-matched interventions confirm that threshold-crossing layers, rather than update magnitude, carry the forgetting, and a spike-budget rule derived from the thresholds, requiring one SVD and no validation sweeps, reduces forgetting by $62\%$ on the most fragile model at no task cost.

cs.LG↗

arXiv2026

Depth-Entropy Guided Sampling for Training-Free LLM Reasoning

Reinforcement learning (RL) has become the dominant paradigm for improving the reasoning capabilities of large language models, but it requires expensive training, curated data, and reward signals. Recent work shows that sampling from sharpened base-model distributions at test time recovers much of the RL gain, yet existing methods rely solely on output-layer likelihoods and ignore the transformer's internal forward-pass dynamics. We introduce Depth-Entropy Guided Sampling (DEGS), a training-free, test-time method that exploits layer-wise entropy collapse as an intrinsic quality signal. We observe that stronger reasoners -- including RL-posttrained variants -- exhibit a distinctive "late collapse": logit-lens decoded entropy stays elevated until deeper layers before converging. We define a per-sequence collapse depth $D(\mathbf{x})$ and a joint objective $\pi(\mathbf{x}) \propto p(\mathbf{x})^\alpha \exp(\beta D(\mathbf{x}))$ that combines sequence likelihood with this depth-entropy structure, instantiated inside an MCMC power-sampling framework (DEGS-MCMC). Across three open-weight models and four reasoning benchmarks, this near-chance per-candidate signal compounds over the sampling trajectory into state-of-the-art training-free accuracy, with gains largest out of domain and on the harder splits -- exactly where likelihood alone falls short -- at single-digit-percent wall-clock overhead. DEGS narrowly trails an in-house GRPO reference on the math splits GRPO was trained for, yet surpasses it out of domain on GPQA for all three models, without any training, reward model, or labeled data.

cs.LG↗

arXiv2026

How Can Reinforcement Learning Achieve Expert-level Placement?

Chip placement is a critical step in physical design. While reinforcement learning (RL)-based methods have recently emerged, their training primarily focuses on wirelength optimization, and therefore often fail to achieve expert-quality layouts. We identify the reward design as the primary cause for the performance gap with experts, and instead of formalizing intricate processes, we circumvent this by directly learning from expert layouts to derive a reward model. Our approach starts from the final expert layouts to infer step-by-step expert trajectories. Using these trajectories as demonstrations or preferences, we train a model that captures the latent implicit rewards in expert results. Experiments show that our framework can efficiently learn from even a single design and generalize well to unseen cases.

cs.AR↗

arXiv2026

FlowPlace: Flow Matching for Chip Placement

Chip placement plays an important role in physical design. While generative models like diffusion models offer promising learning-based solutions, current methods have the following limitations: they use random synthetic data for pre-training, require long sampling times, and often result in overlaps due to their dependence on gradient-based solvers during the sampling process. To overcome these issues, we propose FlowPlace, which features mask-guided synthetic data generation, flow-based efficient training with flexible prior injection, and hard constraint sampling for overlap-free layouts. Experiments on OpenROAD and ICCAD 2015 benchmarks show FlowPlace achieves better PPA metrics, 10-50$\times$ faster sampling efficiency, and zero overlaps.

cs.AR↗

arXiv2026

The Cathaya argyrophylla Genome Reveals the Evolutionary Trade-offs of a Living Fossil

Cathaya argyrophylla is an endangered paleoendemic gymnosperm characterized by restricted ecological adaptability and high pathogen susceptibility. To elucidate its genomic architecture and evolutionary history, a de novo chromosome-level genome assembly was constructed using PacBio High-Fidelity long reads and Hi-C scaffolding. The resulting 22.73 Gb assembly resolves into 12 pseudochromosomes, demonstrating genome gigantism driven primarily by a 72.92 percent repeat sequence content and extensive intron expansion. Phylogenomic analysis using single-copy orthologs identifies C. argyrophylla as a sister lineage to the Pinus clade, with an estimated divergence time of 102.8 million years ago. Analysis of gene family dynamics reveals significant expansions in pathways related to membrane lipid metabolism, transmembrane transport, and translation machinery, indicating specific molecular adaptations for cellular homeostasis in resource-limited environments. Conversely, the genome exhibits massive contractions in endogenous defense networks, including plant-pathogen interactions, brassinosteroid signaling, and DNA repair mechanisms. This distinct genomic reduction correlates directly with the slow growth rate and weak innate immunity observed in the species, while the expanded transmembrane transport networks suggest an obligate physiological reliance on symbiotic microbiomes for survival. Ultimately, this reference genome establishes a critical molecular resource for future conservation and breeding programs.

q-bio.GN↗

arXiv2026

Orthogonal Transformations for Efficient Data-Driven Reachability Analysis

Data-driven reachability analysis using matrix zonotopes faces a fundamental challenge: the number of generators in the reachable set grows exponentially during propagation, while current order reduction yields overly conservative approximations in data-driven settings. This paper introduces an orthogonal matrix-based framework that appropriately transfers the coordinate system before reducing the generators of the reachable set, dramatically reducing reachable set volumes. By exploiting the factorized structure of data-driven matrix zonotope generators, we develop several efficient algorithms to solve the problem. Numerical experiments demonstrate order-of-magnitude volume reductions compared to traditional methods, while maintaining comparable generator numbers. Our method provides a practical solution to improve precision in data-driven safety verification.

eess.SY↗

arXiv2026

Data-Driven Reachability Analysis Using Matrix Perturbation Theory

We propose a matrix zonotope perturbation framework that leverages matrix perturbation theory to characterize how noise-induced distortions alter the dynamics within sets of models. The framework derives interpretable Cai-Zhang bounds for matrix zonotopes (MZs) and extends them to constrained matrix zonotopes (CMZs). Motivated by this analysis and the computational burden of CMZ-based reachable-set propagation, we introduce a coefficient-space approximation in which the constrained coefficient space of the CMZ is over-approximated by an unconstrained zonotope. Replacing CMZ-constrained-zonotope (CZ) products with unconstrained MZ-zonotope multiplication yields a simpler and more scalable reachable-set update. Experimental results demonstrate that the proposed method is substantially faster than the standard CMZ approach while producing reachable sets that are less conservative than those obtained with existing MZ-based methods, advancing practical, accurate, and real-time data-driven reachability analysis.

eess.SY↗

arXiv2026

From Points to Sets: Set-Based Safety Verification in the Latent Space

We extend latent representation methods for safety control design to set-valued states. Recent work has shown that barrier functions designed in a learned latent space can transfer safety guarantees back to the original system, but these methods evaluate certificates at single state points, ignoring state uncertainty. A fixed safety margin can partially address this but cannot adapt to the anisotropic and time-varying nature of the uncertainty gap across different safety constraints. We instead represent the system state as a zonotope, propagate it through the encoder to obtain a latent zonotope, and evaluate certificates over the worst case of the entire set. On a 16-dimensional quadrotor suspended-load gate passage task, set-valued evaluation achieves 5/5 collision-free passages, compared to 1/5 for point-based evaluation and 2/5 for a fixed-margin baseline. Set evaluation reports safety in 44.4% of per-head evaluations versus 48.5% for point-based, and this greater conservatism detects 4.1% blind spots where point evaluation falsely certifies safety, enabling earlier corrective control. The safety gap between point and set evaluation varies up to $12\times$ across certificate heads, explaining why no single fixed margin suffices and confirming the need for per-head, per-timestep adaptation, which set evaluation provides by construction.

eess.SY↗

arXiv2026

Data-Driven Reachability Analysis with Optimal Input Design

This paper addresses the conservatism in data-driven reachability analysis for discrete-time linear systems subject to bounded process noise, where the system matrices are unknown and only input--state trajectory data are available. Building on the constrained matrix zonotope (CMZ) framework, two complementary strategies are proposed to reduce conservatism in reachable-set over-approximations. First, the standard Moore--Penrose pseudoinverse is replaced with a row-norm-minimizing right inverse computed via a second-order cone program (SOCP), which directly reduces the size of the resulting model set, yielding tighter generators and less conservative reachable sets. Second, an online A-optimal input design strategy is introduced to improve the informativeness of the collected data and the conditioning of the resulting model set, thereby reducing uncertainty. The proposed framework extends naturally to piecewise affine systems through mode-dependent data partitioning. Numerical results on a five-dimensional stable LTI system and a two-dimensional piecewise affine system demonstrate that combining designed inputs with the row-norm right inverse significantly reduces conservatism compared to a baseline using random inputs and the pseudoinverse, leading to tighter reachable sets for safety verification.

eess.SY↗

arXiv2026

Bridging Data-Driven Reachability Analysis and Statistical Estimation via Constrained Matrix Convex Generators

Data-driven reachability analysis enables safety verification when first-principles models are unavailable. This requires constructing sets of system models consistent with measured trajectories and noise assumptions. Existing approaches rely on zonotopic or box-based approximations, which do not fit the geometry of common noise distributions such as Gaussian disturbances and can lead to significant conservatism, especially in high-dimensional settings. This paper builds on ellipsotope-based representations to introduce mixed-norm uncertainty sets for data-driven reachability. The highest-density region defines the exact minimum-volume noise confidence set, while Constrained Convex Generators (CCG) and their matrix counterpart (CMCG) provide compatible geometric representations at the noise and parameter level. We show that the resulting CMCG coincides with the maximum-likelihood confidence ellipsoid for Gaussian disturbances, while remaining strictly tighter than constrained matrix zonotopes for mixed bounded-Gaussian noise. For non-convex noise distributions such as Gaussian mixtures, a minimum-volume enclosing ellipsoid provides a tractable convex surrogate. We further prove containment of the CMCG times CCG product and bound the conservatism of the Gaussian-Gaussian interaction. Numerical examples demonstrate substantially tighter reachable sets compared to box-based approximations of Gaussian disturbances. These results enable less conservative safety verification and improve the accuracy of uncertainty-aware control design.

eess.SY↗

arXiv2026

Synonymous Codon Usage Bias Overrides Phylogeny to Reflect Convergent Frond Architecture in a Rapidly Radiating Fern Family Thelypteridaceae

Convergent evolution provides powerful evidence for natural selection, yet its molecular basis is typically sought in protein-coding amino acid substitutions. Whether adaptive pressures can drive the convergent evolution of synonymous codon usage bias (CUB) to override phylogenetic history remains a fundamental question. Here, we investigate this within the rapidly radiating fern family Thelypteridaceae by establishing a comparative framework that integrates chloroplast phylogenomics with dimensionality reduction of codon usage, morphological data, and divergence time estimation. Our results reveal that chloroplast CUB patterns are strikingly incongruent with the phylogeny of this family. Instead, they partition species into distinct clusters that strongly correlate with a convergently evolved morphological trait, lamina base architecture, a key adaptation whose radiation we date to the early Neogene. This convergent molecular signal is driven by a specific subset of photosynthesis-related genes (ndhJ, psaA, and psbD), which exhibit a high density of type-specific, third-position codon substitutions. These findings demonstrate that CUB can serve as a powerful, quantifiable indicator of adaptive history, revealing a cryptic layer of molecular convergence linked to the regulation of protein synthesis. Our work providing a new framework for uncovering adaptive histories obscured by complex evolutionary processes.

q-bio.PE↗

arXiv2026

Transformer-Accelerated Interpolated Data-Driven Reachability Analysis from Noisy Data

Data-driven reachability analysis provides guaranteed outer approximations of reachable sets from input-state measurements, yet each propagation step requires a matrix-zonotope multiplication whose cost grows with the horizon length, limiting scalability. We observe that data-driven propagation is inherently step-size sensitive, in the sense that set-valued operators at different discretization resolutions yield non-equivalent reachable sets at the same physical time, a property absent in model-based propagation. Exploiting this multi-resolution structure, we propose Interpolated Reachability Analysis (IRA), which computes a sparse chain of coarse anchor sets sequentially and reconstructs fine-resolution intermediate sets in parallel across coarse intervals. We derive a fully data-driven coarse-noise over-approximation that removes the need for continuous-time system knowledge, prove deterministic outer-approximation guarantees for all interpolated sets, and establish conditional tightness relative to the fine-resolution chain. To replace the remaining matrix-zonotope multiplications in the fine phase, we further develop Transformer-Accelerated IRA (TA-IRA), where an encoder-decoder Transformer is calibrated via split conformal prediction to provide finite-sample pointwise and path-wise coverage certificates. Numerical experiments on a five-dimensional linear system confirm the theoretical guarantees and demonstrate significant computational savings.

eess.SY↗

arXiv2026

Transformer-Enhanced Data-Driven Output Reachability with Conformal Coverage Guarantees

This paper considers output reachability analysis for linear time-invariant systems with unknown state-space matrices and unknown observation map, given only noisy input-output measurements. The Cayley--Hamilton theorem is applied to eliminate the latent state algebraically, producing an autoregressive input-output model whose parameter uncertainty is enclosed in a matrix zonotope. Set-valued propagation of this model yields output reachable sets with deterministic containment guarantees under a bounded aggregated residual assumption. The conservatism inherent in the lifted matrix-zonotope product is then mitigated by a decoder-only Transformer trained on labels obtained through directional contraction of the formal envelope via an exterior non-reachability certificate. Split conformal prediction restores distribution-free coverage at both per-step and trajectory levels without access to the true reachable-set hull. The framework is validated on a five-dimensional system with multiple unknown observation matrices.

eess.SY↗