arXiv subjects

Hao Wu

Publications and source records attributed to Hao Wu.

At least 91 records · Page 5Linked to original sources

arXiv2026

RAGCell: Retrieval-Augmented Generation as Supervision for Versatile Single-cell Analysis

Single-cell foundation models (scFMs) are transforming computational biology by enabling generalizable, task-agnostic representations for versatile single-cell analysis. Despite their progress in facilitating rapid deployment for downstream tasks, off-the-shelf scFMs still have some overlooked concerns: (I) (Pretraining Cost.) Pretrain-based scFMs necessitate pretraining on a vast volume of cells, rendering it draining resources in applications. (II) (Heterogeneous Gap.) Large Language Models (LLM)-based scFMs ignore the tremendous heterogeneous gap between LLM textual and raw cellular spaces, leading to insufficient capability when facing downstream tasks. To this end, we introduce RAGCell, a versatile single-cell analysis framework that achieves a double-win in both cost-effectiveness and high performance. The success of RAGCell lies in two key aspects: Leveraging LLMs to construct cell-level and feature-level knowledge databases, which serve as supervision signals for training the cell model and significantly reduce the training cost ($>$pretrain-based scFMs). Aligning cell representations with text embeddings from the bi-level knowledge databases, enabling knowledge transfer from textual spaces to cellular spaces and effectively mitigating the heterogeneous gap ($>$LLM-based scFMs). Through extensive experiments on six downstream single-cell analysis tasks, we demonstrate that RAGCell achieves outstanding performance compared to state-of-the-art scFMs while operating at less than $\sim$1/10 the cost of pretrain-based scFMs.

q-bio.GN↗

arXiv2026

Phonon-Bottleneck-Governed Ultrafast Hot-Carrier Super-Diffusion in Transition Metal Dichalcogenides

Two-dimensional transition metal dichalcogenides (TMDCs) are promising for low-power optoelectronics, yet their operational speed is widely considered constrained by low room-temperature mobilities and carrier transit delays. Here, by combining on-chip terahertz optoelectronic sampling with thermally evaporated Ohmic contacts, we eliminate external parasitic delays and directly capture the intrinsic interfacial photoresponse in unencapsulated TMDCs under zero bias. The devices achieve ultrafast relaxation lifetimes of 48.5 ps in MoS2/Au and 14.2 ps in MoSe2/Ag, translating to intrinsic 3-dB bandwidths of 4.4 GHz and 7.5 GHz, respectively. Spatial scanning and bias-dependent measurements show that this response is position-independent and bias-immune, ruling out conventional drift-limited transport and identifying hot-carrier super-diffusion driven by an interfacial electron temperature gradient as the operative mechanism. Furthermore, ultrafast pump-probe spectroscopy reveals that the macroscopic response time is quantitatively synchronized with the microscopic optical-to-acoustic phonon scattering lifetime governed by the intrinsic phonon bottleneck. Our findings establish phonon engineering as a viable paradigm to tailor non-equilibrium optoelectronic dynamics, offering a blueprint for zero-bias, ultrafast, self-powered devices.

cond-mat.mes-hall↗

arXiv2026

UniToolCall: Unifying Tool-Use Representation, Data, and Evaluation for LLM Agents

Tool-use capability is a fundamental component of LLM agents, enabling them to interact with external systems through structured function calls. However, existing research exhibits inconsistent interaction representations, largely overlooks the structural distribution of tool-use trajectories, and relies on incompatible evaluation benchmarks. We present UniToolCall, a unified framework for tool learning that standardizes the entire pipeline from toolset construction and dataset generation to evaluation. The framework curates a large tool pool of 22k+ tools and constructs a hybrid training corpus of 390k+ instances by combining 10 standardized public datasets with structurally controlled synthetic trajectories. It explicitly models diverse interaction patterns, including single-hop vs. multi-hop and single-turn vs. multi-turn, while capturing both serial and parallel execution structures. To support coherent multi-turn reasoning, we further introduce an Anchor Linkage mechanism that enforces cross-turn dependencies. Furthermore, we convert 7 public benchmarks into a unified Query--Action--Observation--Answer (QAOA) representation with fine-grained evaluation at the function-call, turn, and conversation levels. Experiments show that fine-tuning Qwen3-8B on our dataset substantially improves tool-use performance. Under the distractor-heavy Hybrid-20 setting, achieves 93.0% single-turn Strict Precision, outperforming commercial models including GPT, Gemini, and Claude.

cs.AI↗

arXiv2026

Rydberg-atom microwave angle-of-arrival detection via cylindrical vapor-cell-mediated field redistribution

Microwave angle-of-arrival(AoA) measurement is essential for radar, communication, and spectrum mon-itoring. Existing Rydberg-atom-based AoA schemes employ phase-difference measurements with local oscil-lators, standing-wave fluorescence imaging, or amplitude-ratio readout with internal metal reflectors. Here we demonstrate a new approach: a cylindrical glass vapor cell serving directly as an angle-encoding dielectric structure, eliminating the need for multiple apertures, local oscillators, or imaging optics. The cylindrical geometry produces angle-dependent reflection and field redistribution, mapping the incident AoA onto the effective microwave field sampled by the Rydberg ensemble. This effective field is read out optically via the Autler-Townes(A-T) splitting in the electromagnetically induced transparency(EIT) spectrum. Full-wave simulations and experiments at 11.64 GHz confirm a deterministic, geometry-mediated response over 0°-90°.Over the monotonic operating range(55°-125°), the angular resolution(minimum distinguishable increment)is 0.05°, and the angular accuracy(RMSE of repeated measurements) is 0.13° across the full range, improving to 0.05° in the optimal region(65°-115°). Occupying a sensing volume of~2.5 cm3, this method offers a compact, single-sensor pathway toward Rydberg AoA receivers.

physics.optics↗

arXiv2026

SCALE:Scalable Conditional Atlas-Level Endpoint transport for virtual cell perturbation prediction

Virtual-cell models aim to predict how cell populations respond to perturbations, but control and treated cells are measured as unpaired populations, complicating the learning of perturbation-specific effects. We present SCALE, a conditional transport model that represents cells as unordered sets and predicts treated populations without cell-level matching. A shared set-aware encoder and conditional DiT backbone learn latent transport, making endpoint supervision directly delta-aligned without an auxiliary delta objective. Across genetic, chemical, developmental and immune perturbations, SCALE recovered gene-expression changes, response directions and population structure. In CRISPR data with dominant cell-line effects, SCALE outperformed competing methods across seven metrics and maintained separation among gene-target representations rather than collapsing them into a shared region. SCALE further prioritized cytokines predicted to produce distinct immune activation and inflammatory responses. Experiments using matched PBMC samples from three donors confirmed these predicted differences. Together, SCALE enables perturbation-specific prediction from unpaired populations and supports experimental prioritization.

cs.LG↗

arXiv2026

SatDL: Jointly Optimizing Data Redistribution and Training for Satellite-Based Distributed Learning

Satellite-based distributed learning promises to train machine-learning models directly in orbit using massive, globally dispersed sensor data, thereby avoiding large-scale data downloads to ground servers. However, training convergence is significantly slowed by severe non-IID data, specifically label imbalance, as each satellite observes different geographic regions with distinct labels. This imbalance extends training duration and increases energy consumption for solar-powered satellites. Existing approaches either fully redistribute data to enforce IID conditions - accelerating convergence but incurring substantial communication delays - or avoid redistribution entirely by modifying local learning algorithms to mitigate the impact of label imbalance, which, however, still prolong training and increase energy use. Both extremes result in excessive total end-to-end learning time (data-transfer delay plus training time) and thus elevated onboard energy consumption. We present SatDL, a data-redistribution framework designed to minimize total end-to-end learning time. At its core, SatDL develops a Distributor-Critic framework that jointly models and optimizes data-transfer delay and training time. Evaluations through trace-driven simulations of a 1,584-satellite Starlink constellation and hardware emulations using NVIDIA Jetson and A100 GPUs across five datasets show SatDL reduces total end-to-end learning time by up to 18.6% and onboard energy consumption by 12.23-88.00%, while maintaining inference accuracy within a few percentage points of state-of-the-art baselines.

cs.DC↗

arXiv2026

SpectraTac: A Compact Camera-Free Optical Tactile Sensor with Distributed Color Sensing

Tactile sensing is essential for physical interaction in robotics and human--machine systems. However, combining rich tactile information with compact hardware, low cost, and low computational overhead remains challenging. This work presents SpectraTac, a compact, camera-free optical tactile sensor that combines active red--green--blue (RGB) illumination with spatially distributed color sensing. Contact deforms a compliant transparent elastomer and modulates its internal light field, producing spatially differentiated changes in color and intensity. Three distributed color sensors capture these responses as low-dimensional spatio-spectral features, avoiding cameras, imaging optics, and high-dimensional image processing. The device measures 19.2 mm in diameter and 4 mm in height, with a material cost below USD~5. A data-driven decoding framework simultaneously estimates three-dimensional (3D) force and the contact region from the optical measurements. For 3D force prediction, the sensor achieved mean absolute errors (MAEs) of 0.161, 0.164, and 0.429 N along the x-, y-, and z-axes, respectively. The nine-region contact-classification accuracy was 99.9%. We further evaluated real-time 3D force tracking and contact-region-based human--machine interaction through an interactive control task. These results indicate that distributed color-resolved optical sensing offers a compact, low-cost alternative to camera-based tactile sensing for robotics, wearable sensing, and interactive systems.

cs.RO↗

arXiv2026

LARC: Lazy Adaptive Reachability Certification of Robot Manipulator Trajectories

Discrete trajectory checks can miss collisions between sampled robot states. Reachability-based certification bounds motion between states, but uniform time partitions waste computation where clearance is large. We present lazy adaptive reachability certification (LARC), which checks a planned trajectory by bisecting only intervals with an inconclusive clearance test. For piecewise-cubic Hermite joint trajectories, the method bounds link occupancy using midpoint capsules inflated by exact componentwise speed maxima. Certified intervals covering the trajectory provide continuous-time external-obstacle clearance, subject to geometric containment, static obstacles, and a prescribed margin. On 160 AgileX PIPER trajectories from 80 start-goal pairs, LARC matched all decisions of the fixed-fine baseline at depth nine. It used 20328 interval evaluations (24.8% of baseline work), with a median paired speedup of 10.28x. A separate MoveIt/FCL audit checked 158051 states and detected collisions in 21 direct-interpolation controls, none of which LARC certified. The method reduced computation under a shared certificate model, but 27 of 139 sampled-clear trajectories remained uncertified. The sampled audit cannot independently prove continuous-time clearance.

cs.RO↗

arXiv2026

Scoring, Reasoning, and Selecting the Best! Ensembling Large Language Models via a Peer-Review Process

We propose LLM-PeerReview, an unsupervised LLM Ensemble method that selects the most ideal response from multiple LLM-generated candidates for each query, harnessing the collective wisdom of multiple models with diverse strengths. LLM-PeerReview is built on a novel, peer-review-inspired framework that offers a transparent and interpretable mechanism, while remaining fully unsupervised for flexible adaptability and generalization. Specifically, it operates in three stages: For scoring, we use the emerging LLM-as-a-Judge technique to evaluate each response by reusing multiple LLMs at hand; For reasoning, we can apply a straightforward averaging strategy or a principled graphical model-based truth inference algorithm to aggregate multiple scores to produce a final score for each response; Finally, the highest-scoring response is selected as the best ensemble output. LLM-PeerReview is conceptually simple and empirically powerful. Our results across four datasets show that the two variants of the proposed approach outperform the advanced model Smoothie-Global by 6.9% and 7.3% points, cross diverse task types including factual recall QA, math reasoning, and instruction following. Notably, we also establish a carefully curated benchmark suite for LLM Ensemble, integrating 12 methods across four classic datasets and three task families, all evaluated under a rigorous and consistent protocol. We hope this repository will help researchers reproduce the LLM Ensemble baselines.

cs.CL↗

arXiv2026

Beyond FLOPs: Benchmarking Real Inference Acceleration of LLM Pruning under a GEMM-Centric Taxonomy

Pruning has emerged as a dominant paradigm for accelerating large language model (LLM) inference, spanning a broad spectrum of methods that remove computation across tokens, layers, heads, dimensions, and attention patterns. Despite sharing the same objective, these pruning approaches induce fundamentally different execution behaviors, causing realized speedups to depend heavily on hardware and kernel implementations. Consequently, the practical acceleration benefits of different pruning families remain poorly understood. In this work, we introduce a GEMM-centric taxonomy that reorganizes existing pruning methods according to the logical \textbf{M}, \textbf{N}, and \textbf{K} dimensions of general matrix multiplication (GEMM). Leveraging this abstraction, we build a unified benchmarking framework that enables implementation-consistent comparison across the pruning design space and systematically characterizes the acceleration--quality Pareto frontier. Our results on Llama3.1-8B show that static depth pruning remains the strongest Pareto-optimal baseline and stays closest to its theoretical acceleration upper bound in memory-bounded scenarios. During prefill, the frontier transitions from static depth at low quality loss (0\%--4\%), to dynamic depth at moderate loss (5\%--16\%), and finally to static width pruning at higher loss levels (17\%--26\%). These findings establish the first unified view of the practical limits of pruning-based LLM acceleration and provide guidance for future pruning research. Code is available at https://github.com/EIT-NLP/LLM-Pruning/tree/main/PruningInferSim

cs.LG↗

arXiv2026

Relaxation-Aware Multimodal Sensing of Soft Gripper Driven by Structure-Perception-Learning

Achieving stable, sustained grasping with soft robotic hands remains a fundamental challenge. Compliance enables safe and adaptive contact, yet the intrinsic viscoelasticity of soft polymers leads to stress relaxation and a continuous decay of grasping force during holding. Inspired by human grasping, which combines phase-dependent stiffness regulation with continuous sensing and feedback, this paper presents an integrated structure--perception--learning framework. We develop a variable-stiffness soft gripper that uses onboard vision and infrared thermography to track deformation and the temperature field in real time, preserving continuous tracking of the interaction state. To mitigate relaxation-induced force decay, we propose a temperature-coupled viscoelastic force representation, together with a physics-informed learning model, to reconstruct the force trend and provide explicit compensation during holding. Experiments show that, in a 280s force-controlled grasp-and-hold task, the proposed method maintains the desired force with a mean absolute error of 0.066N, outperforming fixed-aperture and instantaneous-only baselines by 80% and 95%, respectively. Overall, the results support a mechanism--AI co-design view: mechanisms shape feasible interactions, while learning compensates remaining uncertainty in viscoelastic dynamics, together enabling stable, sustained grasping.

cs.RO↗

arXiv2026

Active Surface-Driven Reconfigurable Gripper: Robust Grasping and Sequential Manipulation of Thin Objects

Robotic grippers face substantial challenges in grasping and manipulating thin objects. Most existing grippers rely on highly precise approach and grasp motions, which limits robustness and reduces applicability. This paper explores thin-object grasping using books as a representative example. Here, we propose a novel solution that integrates an active surface with underactuated compliance to achieve stable grasping of thin objects without complex control. First, an underactuated gripper with an active surface is designed. The active-surface thumb performs in-hand repositioning of the target book without requiring adjustments of the robot arm or the other fingers, while the underactuated fingers establish compliant contact conditions with the environment, and the reconfigurable structure enables reliable grasping of books under different configurations. Second, we establish a kinematic model of the gripper, and determine the initial grasp postures for two representative scenarios (books lying flat on a desktop and books vertically packed in a shelf). Third, by analyzing the physical model of a book lying on a table and its interaction with the gripper and the environment, we systematically optimize the structural parameters and grasping strategy. Finally, extensive experiments validate the effectiveness of the proposed gripper and strategy. The results demonstrate strong robustness and adaptability when grasping thin objects placed flat (including books, paper, fabric, plastic film, and mouse pad), as well as a high success rate when grasping vertically packed books. Moreover, the proposed gripper can reliably complete long sequential "grasp-place" tasks.

cs.RO↗

arXiv2026

When Context Gets Root: Privilege Escalation in LLM Harnesses

Instruction hierarchy is a model-side defense that assigns instructions different levels of privilege according to their sources. These levels constrain which content may direct model behavior. During agent execution, however, agent harnesses construct context for each model invocation. This construction can elevate low-level content to a higher instruction level and grant it greater model-facing privilege. We introduce instruction privilege escalation. In this attack, an attacker induces an agent to elevate low-level malicious content to a higher instruction level. The elevated content then causes the agent to execute instructions it would not follow at their original level. We evaluate this threat by using multi-agent mechanisms to achieve 13 attack objectives across six coding-agent harnesses. These objectives span confidentiality, integrity, availability, and remote code execution. With unrestricted action execution, the attacks achieve all 13 objectives on all six harnesses. Under automatic permission review, the attacks achieve all 13 objectives on all three harnesses that provide this mode. We further reproduce the vulnerability using harness-provided persistent goals and scheduled tasks. These results demonstrate the generality of instruction privilege escalation.

cs.CR↗

arXiv2026

Global Quasi-strong Solutions to the Voigt Regularization of a Diffuse Interface Model for Incompressible Two-phase Flows with Bulk-surface Interaction

We analyze a thermodynamically consistent diffuse interface model for incompressible two-phase viscous flows with unmatched densities in a smooth bounded domain $Ω\subset\mathbb{R}^d$ $(d=2,3)$. This model consists of the Navier-Stokes-Voigt equations for the velocity field and a convective Cahn-Hilliard equation with a singular potential for the phase-field variable. The resulting hydrodynamic system is subject to a generalized Navier slip boundary condition for the velocity field and Cahn-Hilliard-type dynamic boundary conditions for the phase-field variable and the chemical potential. With the aid of the Voigt regularization, we establish the existence of global quasi-strong solutions that satisfy an energy equality. This is achieved through a combination of delicate approximation schemes and a semi-Galerkin method.

math.AP↗

arXiv2026

Anisotropic Electrostatic-Elastic Softening and Stability in Charged Colloidal Crystals

Charged colloidal crystals exhibit a subtle interplay between electrostatic screening and elastic deformation. When an isotropic electrostatic volumetric softening acts on an anisotropic elastic background, the longitudinal acoustic response softens preferentially along specific crystallographic axes. This article provides a self-contained derivation of the long-wavelength acoustic stability condition for cubic crystals subject to a generic electrostatic-elastic coupling. Starting from an effective static elastic tensor renormalized by a scalar coupling constant $λ_g$, we obtain an explicit condition for the onset of a long-wavelength acoustic instability: the direction $\hat{\mathbf{k}}$ that first loses strong ellipticity is determined by the inverse Christoffel matrix evaluated along that direction. Closed-form expressions for the critical coupling $λ_g^c$ are given for the $[100]$, $[110]$, and $[111]$ high-symmetry directions. We further provide a microscopic derivation of $λ_g$ from the Poisson-Boltzmann theory in a spherical Wigner-Seitz cell, linking the phenomenological constant to experimentally accessible parameters such as salt concentration, particle charge, and volume fraction. The analysis reveals that the most fragile direction can be identified without full lattice-dynamical calculations, and the associated unstable strain patterns are discussed. We also compare these acoustic thresholds with the Born stability condition for homogeneous volume change and show that, for a Born-stable reference cubic crystal, the volume mode becomes unstable before any directional acoustic mode, provided homogeneous dilation is admissible. Numerical illustrations using representative modulus values reported for soft colloidal assemblies illustrate how the criterion can be applied to predict directional softening trends. The present framework serves as a diagnostic tool

cond-mat.soft↗

arXiv2026

Correlation-Aware Structured Pruning for Large Language Models

Structured pruning is a promising approach for reducing the substantial inference costs of Large Language Models (LLMs) while maintaining hardware efficiency. Many existing methods assess the importance of prunable units (e.g., channels or heads) in isolation, implicitly assuming that pruning errors are additive. This independence assumption is often invalidated by the non-orthogonality of model weights and strong correlations between unit activations, potentially leading to performance degradation. To address this, we propose a Correlation-Aware Structured Pruning method. We formulate the pruning objective as a cardinality-constrained binary quadratic program that explicitly models cross-unit dependencies in the reconstruction error. Since this binary quadratic program is NP-hard and difficult to solve exactly, we develop a greedy interaction algorithm based on dependency-aware marginal costs to optimize unit selection. Furthermore, we incorporate a gradient-based strategy to achieve adaptive layer-wise sparsity allocation across the entire model. Extensive experiments on mainstream LLMs demonstrate that incorporating correlation information yields competitive accuracy-efficiency trade-offs compared to representative structured pruning baselines.

cs.CL↗

arXiv2026

$π$-Girsanov: A Generalized Method to Construct Markov State Models from Non-Equilibrium and Multiensemble Biased Simulations

We introduce $π$-Girsanov, a new method for constructing Markov state models from biased enhanced-sampling molecular dynamics simulations based on Girsanov reweighting. The key idea behind this new method is to separate the reweighting of the stationary density from the reweighting of the correlation function. We evaluate the effectiveness of this approach on several analytical potentials and on a model biomolecular system, comparing its performance with the original method. Our results show that $π$-Girsanov not only improves the estimation in a single-ensemble setting, but also resolves key challenges in estimating transition matrices from multiensemble and non-equilibrium biased trajectories. Overall, $π$-Girsanov represents a substantial advance in kinetic reweighting, strengthening the connection between enhanced sampling techniques and Markov state modeling.

physics.bio-ph↗

arXiv2026

Channel Knowledge Map Enabled Low-Complexity Dynamic Radio Environment Reconstruction

Accurate and timely radio environment reconstruction is important but challenging under particularly dynamic transmitter configurations. The conventional methods such as compressed sensing (CS), Kriging method or U-Net typically require environment measurements and reconstruction overhead for radio environment updating as the transmitter locations or radiation patterns change. In this paper, we propose a novel channel knowledge map (CKM)-enabled dynamic radio environment reconstruction method for efficient radio map updating. Specifically, the recently proposed CKM can store reusable path-level propagation knowledge that is decoupled from the transmitter-side radiation characteristics. We can leverage CKM for lightweight forward radio map generation as the transmitter locations and radiation patterns are known, without requiring new target-map measurements. Simulation results show that the proposed method outperforms CS, Kriging, and U-Net in reconstruction accuracy and exhibits strong robustness performance under dynamic transmitter configurations, which demonstrates the potential of the proposed method for flexible and efficient radio environment reconstruction in dynamic wireless networks.

eess.SP↗