arXiv subjects

Siyuan Sun

Publications and source records attributed to Siyuan Sun.

10 recordsLinked to original sources

arXiv2026

Query Expansion Is More Than Generation: Improving Dense Retrieval through Better Integration

Large language models (LLMs) can generate query expansions without task-specific training, yet the same expansions often make a frozen dense retriever worse. We identify an underexplored factor: prior work has often focused on what text is generated, while how generated text is incorporated into dense retrievers has received less systematic attention. By holding generated expansions fixed, we show that performance degradation can often be attributed to the integration method itself. We introduce AnchorQE, a training-free method that separately encodes the original query and its expansion before interpolating them. The interpolation factor is estimated using an unsupervised online strategy that operates over a small part of the unlabeled test stream. Intuitively, our strategy assigns high expansion trust only when expansions are both retrieval-strong and consistent with the original query's retrieved evidence. We show that AnchorQE improves retrieval effectiveness by up to 12.89% when compared to widely-used expansion-only or text-level concatenation baselines across TREC-DL, LoTTE, and BEIR. Further, we show that our online strategy to estimate the interpolation factor outperforms a fixed weight tuned on a development partition by up to 3.81%.

cs.IR↗

arXiv2026

A generic nonparametric value-at-risk estimator for high dimensions

We present in this article a non-parametric value-at-risk (VaR+CVaR) algorithm that remains accurate for an arbitrarily large number of underlying positions. The algorithm solves the two inherent problems of VaR estimation. First, past history is not directly applicable to the future, but all predictions of the future are based on the past. Second, VaR estimation is equivalent to modeling a single corner of a high-dimensional space (the corner where all bets lose simultaneously). The algorithm only uses mathematical methods that strictly do not degrade in accuracy at high-dimensions. Historical data are then directly incorporated with all high-dimensional relationships present, without manipulation. We test the algorithm with an ensemble of 500 portfolios with random positions across 49 distinct liquid futures of different expiries (VIX, equity indexes, gov. bonds, rates, energy, metals, livestock, agriculture, and softs). All VaR estimations are performed strictly blind to the future. The median portfolio rate of loss exceeding the 99% confidence daily VaR estimate is between $1.0\pm0.1$% depending on algorithm input parameters. 68% of portfolios have a rate of loss exceeding 99% VaR between $1.0\pm0.3$%, and 95% of portfolios between $1.0\pm0.5$%.

q-fin.RM↗

arXiv2026

KernelScript: Cross-Boundary Typed DSL for eBPF Applications

eBPF lets developers extend Linux with custom packet processing, tracing, and scheduling logic, and a verifier proves before execution that the code will not crash the kernel. The programming model, however, is fragmented: a single application spans kernel code, a userspace loader, and shared maps, yet the relationships among these pieces go unchecked. E.g. A map or event type defined differently on each side silently corrupts shared state. We observe that these cross-boundary relationships duplicate information that a type system can unify. We present KernelScript, a DSL that types maps, program handles, and execution domains in one source, then compiles to standard C through the original toolchain. We evaluate KernelScript on 43 eBPF workloads covering XDP, TC, kprobe, tracepoint, and struct_ops. KernelScript rejects cross-boundary bugs at compile time that standard C/libbpf still builds and loads, a unified source shrinks the diffs for cross-boundary changes by 5x, and generated code remains compatible with the existing toolchain.

cs.PL↗

arXiv2026

Creating a biologically more accurate spider robot to study active vibration sensing

Orb-weaving spiders detect prey on a web using vibration sensors at leg joints. They often dynamically crouch their legs during prey sensing, likely an active sensing strategy. However, how leg crouching enhances sensing is poorly understood, because measuring system vibrations in behaving animals is difficult. We use robophysical modeling to study this problem. Our previous spider robot had only four legs, simplified leg morphology, and a shallow crouching range of motion. Here, we developed a new spider robot, with eight legs, each with four joints that better approximated spider leg morphology. Leg exoskeletons were 3-D printed and joint stiffness was tuned using integrated silicone molding with variable materials and geometry. Tendon-driven actuation allowed a motor in the body to crouch all eight legs deeply as spiders do, while accelerometers at leg joints record leg vibrations. Experiments showed that our new spider robot reproduced key vibration features observed in the previous robot while improving biological accuracy. Our new robot provides a biologically more accurate robophysical model for studying how leg behaviors modulate vibration sensing on a web.

cs.RO↗

arXiv2025

A multi-weight self-matching visual explanation for cnns on sar images

In recent years, convolutional neural networks (CNNs) have achieved significant success in various synthetic aperture radar (SAR) tasks. However, the complexity and opacity of their internal mechanisms hinder the fulfillment of high-reliability requirements, thereby limiting their application in SAR. Improving the interpretability of CNNs is thus of great importance for their development and deployment in SAR. In this paper, a visual explanation method termed multi-weight self-matching class activation mapping (MS-CAM) is proposed. MS-CAM matches SAR images with the feature maps and corresponding gradients extracted by the CNN, and combines both channel-wise and element-wise weights to visualize the decision basis learned by the model in SAR images. Extensive experiments conducted on a self-constructed SAR target classification dataset demonstrate that MS-CAM more accurately highlights the network's regions of interest and captures detailed target feature information, thereby enhancing network interpretability. Furthermore, the feasibility of applying MS-CAM to weakly-supervised obiect localization is validated. Key factors affecting localization accuracy, such as pixel thresholds, are analyzed in depth to inform future work.

cs.CV↗

arXiv2025

Global Convergence in Neural ODEs: Impact of Activation Functions

Neural Ordinary Differential Equations (ODEs) have been successful in various applications due to their continuous nature and parameter-sharing efficiency. However, these unique characteristics also introduce challenges in training, particularly with respect to gradient computation accuracy and convergence analysis. In this paper, we address these challenges by investigating the impact of activation functions. We demonstrate that the properties of activation functions, specifically smoothness and nonlinearity, are critical to the training dynamics. Smooth activation functions guarantee globally unique solutions for both forward and backward ODEs, while sufficient nonlinearity is essential for maintaining the spectral properties of the Neural Tangent Kernel (NTK) during training. Together, these properties enable us to establish the global convergence of Neural ODEs under gradient descent in overparameterized regimes. Our theoretical findings are validated by numerical experiments, which not only support our analysis but also provide practical guidelines for scaling Neural ODEs, potentially leading to faster training and improved performance in real-world applications.

cs.LG↗

arXiv2023

High Rate Studies of the ATLAS sTGC Detector and Optimization of the Filter Circuit on the Input of the Front-End Amplifier

The Large Hadron Collider (LHC) at CERN is expected to be upgraded to the High-Luminosity LHC (HL-LHC) by 2029 and achieve instantaneous luminosity around 5 - 7.5 $\times$ 10$^{34}$cm$^{-2}$ s$^{-1}$. This represents a more than 3-4 fold increase in the instantaneous luminosity compared to what has been achieved in Run 2. The New Small Wheel (NSW) upgrade is designed to be able to operate efficiently in this high background rate environment. In this article, we summarize multiple performance studies of the small-strip Thin Gap Chamber (sTGC) at high rate using nearly final front-end electronics. We demonstrate that the efficiency versus rate distribution can be well described by an exponential decay with electronics dead-time being the primary cause of loss of efficiency at high rate. We then demonstrate several methods that can decrease the electronics dead-time and therefore minimize efficiency loss. One such method is to install either a pi-network input filter or pull-up resistor to minimize the charge input into the amplifier. We optimized the pi-network capacitance and pull-up resistor resistance using the results from our measurements. The results shown here were not only critical to finalizing the components on the front-end board, but also are critical for setting the optimal operating parameters of the sTGC detector and electronics in the ATLAS cavern.

physics.ins-det↗

arXiv2021

Design and testing of an sTGC ASIC interface board for the ATLAS New Small Wheel upgrade

The ATLAS experiment will replace the present Small Wheel (SW) detector with a New Small Wheel detector (NSW) aiming to improve the performance of muon triggering and precision tracking in the endcap region at the High-Luminosity LHC. Small-strip Thin Gap Chamber (sTGC) is one of the two new detector technologies used in this upgrade. A few custom-designed ASICs are needed for the sTGC detector. We designed an sTGC ASIC interface board to test ASIC-to-ASIC communication and validate the functionality of the entire system. A test platform with the final readout system is set up and the whole sTGC readout chain is demonstrated for the first time. Key parameters in the readout chain are discussed and the results are shown.

physics.ins-det↗

arXiv2020

Towards Reconfigurable Intelligent Surfaces Powered Green Wireless Networks

The adoption of reconfigurable intelligent surface (RIS) in wireless networks can enhance the spectrum- and energy-efficiency by controlling the propagation environment. Although the RIS does not consume any transmit power, the circuit power of the RIS cannot be ignored, especially when the number of reflecting elements is large. In this paper, we propose the joint design of beamforming vectors at the base station, active RIS set, and phase-shift matrices at the active RISs to minimize the network power consumption, including the RIS circuit power consumption, while taking into account each user's target data rate requirement and each reflecting element's constant modulus constraint. However, the formulated problem is a mixed-integer quadratic programming (MIQP) problem, which is NP-hard. To this end, we present an alternating optimization method, which alternately solves second order cone programming (SOCP) and MIQP problems to update the optimization variables. Specifically, the MIQP problem is further transformed into a semidefinite programming problem by applying binary relaxation and semidefinite relaxation. Finally, an efficient algorithm is developed to solve the problem. Simulation results show that the proposed algorithm significantly reduces the network power consumption and reveal the importance of taking into account the RIS circuit power consumption.

cs.IT↗

arXiv2020

Adversarial Multimodal Network for Movie Question Answering

Visual question answering by using information from multiple modalities has attracted more and more attention in recent years. However, it is a very challenging task, as the visual content and natural language have quite different statistical properties. In this work, we present a method called Adversarial Multimodal Network (AMN) to better understand video stories for question answering. In AMN, as inspired by generative adversarial networks, we propose to learn multimodal feature representations by finding a more coherent subspace for video clips and the corresponding texts (e.g., subtitles and questions). Moreover, we introduce a self-attention mechanism to enforce the so-called consistency constraints in order to preserve the self-correlation of visual cues of the original video clips in the learned multimodal representations. Extensive experiments on the MovieQA dataset show the effectiveness of our proposed AMN over other published state-of-the-art methods.

cs.CV↗