arXiv subjects

Hao Zeng

Publications and source records attributed to Hao Zeng.

At least 37 records · Page 2Linked to original sources

arXiv2026

On the Provable Performance Guarantee of Efficient Reasoning Models

Large reasoning models (LRMs) have achieved remarkable progress in complex problem-solving tasks. Despite this success, LRMs typically suffer from high computational costs during deployment, highlighting a need for efficient inference. A practical direction of efficiency improvement is to switch the LRM between thinking and non-thinking modes dynamically. However, such approaches often introduce additional reasoning errors and lack statistical guarantees for the performance loss, which are critical for high-stakes applications. In this work, we propose Probably Approximately Correct (PAC) reasoning that controls the performance loss under the user-specified tolerance. Specifically, we construct an upper confidence bound on the performance loss and determine a threshold for switching to the non-thinking model. Theoretically, using the threshold to switch between the thinking and non-thinking modes ensures bounded performance loss in a distribution-free manner. Our comprehensive experiments on reasoning benchmarks show that the proposed method can save computational budgets and control the user-specified performance loss.

cs.AI↗

arXiv2026

Conditional Performance Guarantee for Large Reasoning Models

Large reasoning models have shown strong performance through extended chain-of-thought reasoning, yet their computational cost remains significant. Probably approximately correct (PAC) reasoning provides statistical guarantees for efficient reasoning by adaptively switching between thinking and non-thinking models, but the guarantee holds only in the marginal case and does not provide exact conditional coverage. We propose G-PAC reasoning, a practical framework that provides PAC-style guarantees at the group level by partitioning the input space. We develop two instantiations: Group PAC (G-PAC) reasoning for known group structures and Clustered PAC (C-PAC) reasoning for unknown groupings. We prove that both G-PAC and C-PAC achieve group-conditional risk control, and that grouping can strictly improve efficiency over marginal PAC reasoning in heterogeneous settings. Our experiments on diverse reasoning benchmarks demonstrate that G-PAC and C-PAC successfully achieve group-conditional risk control while maintaining substantial computational savings.

cs.AI↗

arXiv2026

Distribution-informed Efficient Conformal Prediction for Full Ranking

Quantifying uncertainty is critical for the safe deployment of ranking models in real-world applications. Recent work offers a rigorous solution using conformal prediction in a full ranking scenario, which aims to construct prediction sets for the absolute ranks of test items based on the relative ranks of calibration items. However, relying on upper bounds of non-conformity scores renders the method overly conservative, resulting in substantially large prediction sets. To address this, we propose Distribution-informed Conformal Ranking (DCR), which produces efficient prediction sets by deriving the exact distribution of non-conformity scores. In particular, we find that the absolute ranks of calibration items follow Negative Hypergeometric distributions, conditional on their relative ranks. DCR thus uses the rank distribution to derive non-conformity score distribution and determine conformal thresholds. We provide theoretical guarantees that DCR achieves improved efficiency over the baseline while ensuring valid coverage under mild assumptions. Extensive experiments demonstrate the superiority of DCR, reducing average prediction set size by up to 36%, while maintaining valid coverage.

cs.LG↗

arXiv2026

HyPAC: Cost-Efficient LLMs-Human Hybrid Annotation with PAC Error Guarantees

Data annotation often involves multiple sources with different cost-quality trade-offs, such as fast large language models (LLMs), slow reasoning models, and human experts. In this work, we study the problem of routing inputs to the most cost-efficient annotation source while controlling the labeling error on test instances. We propose \textbf{HyPAC}, a method that adaptively labels inputs to the most cost-efficient annotation source while providing distribution-free guarantees on annotation error. HyPAC calibrates two decision thresholds using importance sampling and upper confidence bounds, partitioning inputs into three regions based on uncertainty and routing each to the appropriate annotation source. We prove that HyPAC achieves the minimum expected cost with a probably approximately correct (PAC) guarantee on the annotation error, free of data distribution and pre-trained models. Experiments on common benchmarks demonstrate the effectiveness of our method, reducing the annotation cost by 78.51\% while tightly controlling the annotation error.

cs.LG↗

arXiv2025

SQ-format: A Unified Sparse-Quantized Hardware-friendly Data Format for LLMs

Post-training quantization (PTQ) plays a crucial role in the democratization of large language models (LLMs). However, existing low-bit quantization and sparsification techniques are difficult to balance accuracy and efficiency due to the limited hardware support. For example, W4A8 can only achieve the same peak TOPS as W8A8 whereas the GPU-supported sparse data format (2:4 semi-structure sparse) is seldomly adopted due to the loss of accuracy. To bridge this gap, in this paper, we propose the Sparse-Quantized Format (SQ-format), which is a unified data format for quantization and sparsification potentially easily supported by new hardware and existing GPUs. SQ-format makes use of the fact that sparse matrix can be accelerated in high-precision, and low-precision matrix multiplication can also be accelerated accordingly. As such, SQ-format is proposed to achieve Pareto improvement between performance and throughput. This format is particularly suitable for activations with outlier inequality status and makes their static compression possible. We show the state-of-the-art PTQ performance with SQ-format, propose the hardware required to support it, and further offer the design exploration and insights for the next-generation AI accelerators.

cs.CL↗

arXiv2025

Marangoni-driven freezing dynamics of supercooled binary droplets

Solidification of droplets is of great importance to various technological applications, drawing considerable attention from scientists aiming to unravel the fundamental physical mechanisms. In the case of multicomponent droplets undergoing solidification, the emergence of concentration gradients may trigger significant interfacial flows that dominate the freezing dynamics. Here, we experimentally investigate the fascinating interfacial freezing dynamics of supercooled ethanol-water droplets, accompanied with the migration and growth of massive ice particles. We reveal that these unique freezing dynamics are driven by solidification-induced solutal Marangoni flow within the droplets. Our model, which incorporates the temperature- and concentration-dependent properties of the ethanol-water mixture, quantitatively predicts both the migration velocity and the growth rate of the ice particles. The former is determined by the solutal Marangoni flow velocity, while the latter is governed by a balance between the latent heat release and the enhanced thermal dissipation by the Marangoni flow. Moreover, we show that the final wrapping state of droplets can be modulated by the concentration of ethanol. Our findings may pave the way for novel insights into the physicochemical hydrodynamics of multicomponent liquids undergoing phase transitions.

physics.flu-dyn↗

arXiv2025

A Novel Spreading-Factor-Index-Aided LoRa Scheme: Design and Performance Analysis

LoRa is a widely recognized modulation technology in the field of low power wide area networks (LPWANs). However, the data rate of LoRa is too low to satisfy the requirements of Internet of Things applications. To address this issue, we propose a novel high-data-rate LoRa scheme based on the spreading factor index (SFI). In the proposed SFI-LoRa scheme, the starting frequency bin of a chirp signal is used to transmit information bits, while the combinations of spreading factors are exploited as a set of indices to convey additional information bits. Moreover, the theoretical symbol error rate, data rate, transmission throughput, complexity and energy efficiency of the proposed SFI-LoRa scheme are carefully analyzed. Simulation results not only verify the accuracy of our theoretical analysis, but also demonstrate that the proposed SFI-LoRa scheme can improve the transmission throughput of existing LoRa schemes without sacrificing the BER performance over additive white Gaussian noise, Rayleigh fading, and multipath flat-fading channels. Therefore, the proposed SFI-LoRa scheme is a potential solution for applications requiring a high data rate in the LPWAN domain.

eess.SP↗

arXiv2025

Exploring the Noise Robustness of Online Conformal Prediction

Conformal prediction is an emerging technique for uncertainty quantification that constructs prediction sets guaranteed to contain the true label with a predefined probability. Recent work develops online conformal prediction methods that adaptively construct prediction sets to accommodate distribution shifts. However, existing algorithms typically assume perfect label accuracy which rarely holds in practice. In this work, we investigate the robustness of online conformal prediction under uniform label noise with a known noise rate, in both constant and dynamic learning rate schedules. We show that label noise causes a persistent gap between the actual mis-coverage rate and the desired rate $α$, leading to either overestimated or underestimated coverage guarantees. To address this issue, we propose Noise Robust Online Conformal Prediction (dubbed NR-OCP) by updating the threshold with a novel robust pinball loss, which provides an unbiased estimate of clean pinball loss without requiring ground-truth labels. Our theoretical analysis shows that NR-OCP eliminates the coverage gap in both constant and dynamic learning rate schedules, achieving a convergence rate of $\mathcal{O}(T^{-1/2})$ for both empirical and expected coverage errors under uniform label noise. Extensive experiments demonstrate the effectiveness of our method by achieving both precise coverage and improved efficiency.

cs.LG↗

arXiv2025

Agile manoeuvring of dandelion-inspired micro-flyers with vortex-enabled stability

Manoeuvring untethered, centimetre-scale airborne structures has been a long-standing challenge. Active flight systems, relying on high-power-density actuators alongside mechanical and electronic components, are constrained by critical limitations in energy delivery and miniaturisation. In contrast, passive systems transported and distributed by the wind typically lack the capability for mid-air controlled manoeuvrability. Here we report an ultra-light (1.2 mg) hexagonal polymeric assembly capable of passive flight with optical control of its trajectory. This dandelion-inspired micro-flyer incorporates six radially arranged filamentous structures, of which morphology is dynamically controlled through photomechanical deformation by six independent soft actuators made of liquid crystalline elastomer thin films. Compared to the diaspore of the dandelion (Taraxacum officinale), micro-flyer demonstrate a similar terminal velocity (~0.5 m s-1), 45% better positional stability and nearly zero rotational rate (1.68 s-1; natural seeds: 50.8 s-1). Particle image velocimetry reveals that a stable asymmetric separated vortex ring underlies its flight stability, enabling mid-air steerability. When free-falling in a low-turbulent airstream, the light-driven hexapodal fliers demonstrate precise altitude control, reversible body flipping, pattern formation, interactive swarm, and controlled trajectories across three-dimensional space. The results show that responsive materials with light-induced asymmetry can bring about manoeuvrability in air, paving the way for agile, untethered controlled micro-fliers.

physics.app-ph↗

arXiv2025

GLM-4.5: Agentic, Reasoning, and Coding (ARC) Foundation Models

We present GLM-4.5, an open-source Mixture-of-Experts (MoE) large language model with 355B total parameters and 32B activated parameters, featuring a hybrid reasoning method that supports both thinking and direct response modes. Through multi-stage training on 23T tokens and comprehensive post-training with expert model iteration and reinforcement learning, GLM-4.5 achieves strong performance across agentic, reasoning, and coding (ARC) tasks, scoring 70.1% on TAU-Bench, 91.0% on AIME 24, and 64.2% on SWE-bench Verified. With much fewer parameters than several competitors, GLM-4.5 ranks 3rd overall among all evaluated models and 2nd on agentic benchmarks. We release both GLM-4.5 (355B parameters) and a compact version, GLM-4.5-Air (106B parameters), to advance research in reasoning and agentic AI systems. Code, models, and more information are available at https://github.com/zai-org/GLM-4.5.

cs.CL↗

arXiv2025

FFGAF-SNN: The Forward-Forward Based Gradient Approximation Free Training Framework for Spiking Neural Networks

Spiking Neural Networks (SNNs) offer a biologically plausible framework for energy-efficient neuromorphic computing. However, it is a challenge to train SNNs due to their non-differentiability, efficiently. Existing gradient approximation approaches frequently sacrifice accuracy and face deployment limitations on edge devices due to the substantial computational requirements of backpropagation. To address these challenges, we propose a Forward-Forward (FF) based gradient approximation-free training framework for Spiking Neural Networks, which treats spiking activations as black-box modules, thereby eliminating the need for gradient approximation while significantly reducing computational complexity. Furthermore, we introduce a class-aware complexity adaptation mechanism that dynamically optimizes the loss function based on inter-class difficulty metrics, enabling efficient allocation of network resources across different categories. Experimental results demonstrate that our proposed training framework achieves test accuracies of 99.58%, 92.13%, and 75.64% on the MNIST, Fashion-MNIST, and CIFAR-10 datasets, respectively, surpassing all existing FF-based SNN approaches. Additionally, our proposed method exhibits significant advantages in terms of memory access and computational power consumption.

cs.CV↗

arXiv2025

Parametric Scaling Law of Tuning Bias in Conformal Prediction

Conformal prediction is a popular framework of uncertainty quantification that constructs prediction sets with coverage guarantees. To uphold the exchangeability assumption, many conformal prediction methods necessitate an additional holdout set for parameter tuning. Yet, the impact of violating this principle on coverage remains underexplored, making it ambiguous in practical applications. In this work, we empirically find that the tuning bias - the coverage gap introduced by leveraging the same dataset for tuning and calibration, is negligible for simple parameter tuning in many conformal prediction methods. In particular, we observe the scaling law of the tuning bias: this bias increases with parameter space complexity and decreases with calibration set size. Formally, we establish a theoretical framework to quantify the tuning bias and provide rigorous proof for the scaling law of the tuning bias by deriving its upper bound. In the end, we discuss how to reduce the tuning bias, guided by the theories we developed.

cs.LG↗

arXiv2025

Additive Manufacturing of Lunar Regolith for Reconfigurable Building Blocks toward Lunar Habitation

Utilizing locally available materials is a crucial step towards sustainable planetary habitation. Lunar regolith has gained tremendous interest in additive manufacturing in the past decades. However, due to the constrained manufacturing facilities and materials on the moon, many existing additive manufacturing methods are not suitable for practical on-site manufacturing. Here, we envision that light-based direct sintering of lunar regolith can be a feasible approach. Instead of directly manufacturing large structures, we hypothesize that small-scale, reconfigurable building blocks can be an alternative to form large and complex structures. To verify the feasibility, we conducted laser sintering of lunar regolith simulants as a proof of concept, following a simple theoretical calculation for direct sintering using the light available in space. Different laser processing parameters are investigated to obtain controllable lunar regolith sintering. We further designed Lego-like interlocking bricks that are reconfigurable for different structure assemblies without additional material. Mechanical performance (compressive strength) of sintered cubic blocks is evaluated, showing a peak stress of ~1.5 MPa. We hope this work will inspire other in-space manufacturing techniques and enable low-cost space habitation.

astro-ph.IM↗

arXiv2025

Exploring Imbalanced Annotations for Effective In-Context Learning

Large language models (LLMs) have shown impressive performance on downstream tasks through in-context learning (ICL), which heavily relies on the demonstrations selected from annotated datasets. However, these datasets often exhibit long-tailed class distributions in real-world scenarios, leading to biased demonstration selection. In this work, we show that such class imbalances significantly degrade the ICL performance across various tasks, regardless of selection methods. Moreover, classical rebalancing methods, which focus solely on class weights, yield poor performance due to neglecting condition bias--skewed feature distributions within classes. To address this, we propose Reweighting with Conditional Bias (dubbed RCB), a simple and complementary approach to enhance ICL performance under class imbalance. In particular, RCB estimates conditional bias using a balanced subset and re-weights demonstration scores based on both class weight and conditional bias. In effect, RCB prevents over-selection from dominant classes while preserving the efficacy of current selection methods. Extensive experiments on common benchmarks demonstrate the effectiveness of our method, improving the average accuracy of current selection methods by up to 5.42%.

cs.CL↗

arXiv2025

C-Adapter: Adapting Deep Classifiers for Efficient Conformal Prediction Sets

Conformal prediction, as an emerging uncertainty quantification technique, typically functions as post-hoc processing for the outputs of trained classifiers. To optimize the classifier for maximum predictive efficiency, Conformal Training rectifies the training objective with a regularization that minimizes the average prediction set size at a specific error rate. However, the regularization term inevitably deteriorates the classification accuracy and leads to suboptimal efficiency of conformal predictors. To address this issue, we introduce \textbf{Conformal Adapter} (C-Adapter), an adapter-based tuning method to enhance the efficiency of conformal predictors without sacrificing accuracy. In particular, we implement the adapter as a class of intra order-preserving functions and tune it with our proposed loss that maximizes the discriminability of non-conformity scores between correctly and randomly matched data-label pairs. Using C-Adapter, the model tends to produce extremely high non-conformity scores for incorrect labels, thereby enhancing the efficiency of prediction sets across different coverage rates. Extensive experiments demonstrate that C-Adapter can effectively adapt various classifiers for efficient prediction sets, as well as enhance the conformal training method.

cs.LG↗

arXiv2025

Beam test result and digitization of TaichuPix-3: A Monolithic Active Pixel Sensors for CEPC vertex detector

The Circular Electron-Positron Collider (CEPC), as the next-generation electron-positron collider, is tasked with advancing not only Higgs physics but also the discovery of new physics. Achieving these goals requires high-precision measurements of particles. Taichu seires, Monolithic Active Pixel Sensor (MAPS), a key component of the vertex detector for CEPC was designed to meet the CEPC's requirements. For the geometry of vertex detector is long barrel with no endcap, and current silicon lacks a complete digitization model, precise estimation of cluster size particularly causing by particle with large incident angle is needed. Testbeam results were conducted at the Beijing Synchrotron Radiation Facility (BSRF) to evaluate cluster size dependence on different incident angles and threshold settings. Experimental results confirmed that cluster size increases with incident angle. Simulations using the Allpix$^2$ framework replicated experimental trends at small angles but exhibited discrepancies at large angles, suggesting limitations in linear electric field assumptions and sensor thickness approximations. The results from both testbeam and simulations have provided insights into the performance of the TaichuPix chip at large incident angles, offering a crucial foundation for the establishment of a digital model and addressing the estimation of cluster size in the forward region of the long barrel. Furthermore, it offers valuable references for future iterations of TaichuPix, the development of digital models, and the simulation and estimation of the vertex detector's performance.

physics.ins-det↗

arXiv2025

Light communicative materials

The natural interactive materials under far-from-equilibrium conditions have significantly inspired advances in synthetic biomimetic materials. In artificial systems, gradient diffusion serves as the primary means of interaction between individuals, lacking directionality, sufficient interaction ranges and transmission rates. Here, we present a method for constructing highly directed, communicative structures via optical feedback in light responsive materials. We showcase a photomechanical operator system comprising a baffle and a soft actuator. Positive and negative operators are configured to induce light-triggered deformations, alternately interrupting the passage of two light beams in a closed feedback loop. The fundamental functionalities of this optically interconnected material loop include homeostasis-like self-oscillation and signal transmission from one material to another via light. Refinements in alignment facilitate remote sensing, fiber-optic/long-distance communication, and adaptation. These proof-of-concept demonstrations outline a versatile design framework for light-mediated communication among responsive materials, with broad applicability across diverse materials.

physics.optics↗

arXiv2024

Feedback Regulated Opto-Mechanical Soft Robotic Actuators

Natural organisms can convert environmental stimuli into sensory feedback to regulate their body and realize active adaptivity. However, realizing such a feedback-regulation mechanism in synthetic material systems remains a grand challenge. It is believed that achieving complex feedback mechanisms in responsive materials will pave the way toward autonomous, intelligent structure and actuation without complex electronics. Inspired by living systems, we report a general principle to design and construct such feedback loops in light-responsive materials. Specifically, we design a baffle-actuator mechanism to incorporate programmed feedback into the opto-mechanical responsiveness. By simply addressing the baffle position with respect to the incident light beam, positive and negative feedback are programmed. We demonstrate the transformation of a light-bending strip into a switcher, where the intensity of light determines the energy barrier under positive feedback, realizing multi-stable shape-morphing. By leveraging the negative feedback and associated homeostasis, we demonstrate two soft robots, i.e., a locomotor and a swimmer. Furthermore, we unveil the ubiquity of feedback in light-responsive materials, which provides new insight into self-regulated robotic matters.

cs.RO↗