arXiv subjects

Yu Xu

Publications and source records attributed to Yu Xu.

At least 73 records · Page 4Linked to original sources

arXiv2024

Enhancing and Assessing Instruction-Following with Fine-Grained Instruction Variants

The effective alignment of Large Language Models (LLMs) with precise instructions is essential for their application in diverse real-world scenarios. Current methods focus on enhancing the diversity and complexity of training and evaluation samples, yet they fall short in accurately assessing LLMs' ability to follow similar instruction variants. We introduce an effective data augmentation technique DeMoRecon that decomposes complex instructions into simpler sub-components, modifies these, and reconstructs them into new variants, thereby preserves the original instruction's context and complexity while introducing variability, which is critical for training and evaluating LLMs' instruction-following precision. Based on DeMoRecon, we developed the FGIV dataset which contains fine-grained instruction variants of 1,773 seed instructions to both fine-tune and evaluate LLMs. Our findings show that LLMs fine-tuned with FGIV will gain significant performance boost on both ours and commonly used instructions-following benchmarks.

cs.AI↗

arXiv2024

TaCIE: Enhancing Instruction Comprehension in Large Language Models through Task-Centred Instruction Evolution

Large Language Models (LLMs) require precise alignment with complex instructions to optimize their performance in real-world applications. As the demand for refined instruction tuning data increases, traditional methods that evolve simple seed instructions often struggle to effectively enhance complexity or manage difficulty scaling across various domains. Our innovative approach, Task-Centered Instruction Evolution (TaCIE), addresses these shortcomings by redefining instruction evolution from merely evolving seed instructions to a more dynamic and comprehensive combination of elements. TaCIE starts by deconstructing complex instructions into their fundamental components. It then generates and integrates new elements with the original ones, reassembling them into more sophisticated instructions that progressively increase in difficulty, diversity, and complexity. Applied across multiple domains, LLMs fine-tuned with these evolved instructions have substantially outperformed those tuned with conventional methods, marking a significant advancement in instruction-based model fine-tuning.

cs.CY↗

arXiv2024

Cryogenic microwave performance of silicon nitride and amorphous silicon deposited using low-temperature ICPCVD

Fabrication of dielectrics at low temperature is required for temperature-sensitive detectors. For superconducting detectors, such as transition edge sensors and kinetic inductance detectors, AlMn is widely studied due to its variable superconducting transition temperature at different baking temperatures. Experimentally only the highest baking temperature determines AlMn transition temperature, so we need to control the wafer temperature during the whole process. In general, the highest process temperature happens during dielectric fabrication. Here, we present the cryogenic microwave performance of Si$_{3}$N$_{4}$, SiN$_{x}$ and $α$-Si using ICPCVD at low temperature of 75 $^{\circ}$C. The dielectric constant, internal quality factor and TLS properties are studied using Al parallel plate resonators.

physics.ins-det↗

arXiv2024

Instruction Fusion: Advancing Prompt Evolution through Hybridization

The fine-tuning of Large Language Models (LLMs) specialized in code generation has seen notable advancements through the use of open-domain coding queries. Despite the successes, existing methodologies like Evol-Instruct encounter performance limitations, impeding further enhancements in code generation tasks. This paper examines the constraints of existing prompt evolution techniques and introduces a novel approach, Instruction Fusion (IF). IF innovatively combines two distinct prompts through a hybridization process, thereby enhancing the evolution of training prompts for code LLMs. Our experimental results reveal that the proposed novel method effectively addresses the shortcomings of prior methods, significantly improving the performance of Code LLMs across five code generation benchmarks, namely HumanEval, HumanEval+, MBPP, MBPP+ and MultiPL-E, which underscore the effectiveness of Instruction Fusion in advancing the capabilities of LLMs in code generation.

cs.AI↗

arXiv2024

Experimental quantum computational chemistry with optimised unitary coupled cluster ansatz

Quantum computational chemistry has emerged as an important application of quantum computing. Hybrid quantum-classical computing methods, such as variational quantum eigensolvers (VQE), have been designed as promising solutions to quantum chemistry problems, yet challenges due to theoretical complexity and experimental imperfections hinder progress in achieving reliable and accurate results. Experimental works for solving electronic structures are consequently still restricted to nonscalable (hardware efficient) or classically simulable (Hartree-Fock) ansatz, or limited to a few qubits with large errors. The experimental realisation of scalable and high-precision quantum chemistry simulation remains elusive. Here, we address the critical challenges {associated with} solving molecular electronic structures using noisy quantum processors. Our protocol presents significant improvements in the circuit depth and running time, key metrics for chemistry simulation. Through systematic hardware enhancements and the integration of error mitigation techniques, we push forward the limit of experimental quantum computational chemistry and successfully scale up the implementation of VQE with an optimised unitary coupled-cluster ansatz to 12 qubits. We produce high-precision results of the ground-state energy for molecules with error suppression by around two orders of magnitude. We achieve chemical accuracy for H$_2$ at all bond distances and LiH at small bond distances in the experiment, even beyond the two recent concurrent works. Our work demonstrates a feasible path towards a scalable solution to electronic structure calculation, validating the key technological features and identifying future challenges for this goal.

quant-ph↗

arXiv2024

Simulated Coronal Mass Ejections on a young Solar-Type Star and the Associated Instantaneous Angular Momentum Loss

Coronal mass ejections (CMEs) on stars can change the stars' magnetic field configurations and mass loss rates during the eruption and propagation and therefore, may affect the stars' rotation properties on long time-scales. The dynamics of stellar CMEs and their influence on the stellar angular momentum loss rate are not yet well understood. In order to start investigating these CME-related aspects on other stars, we conducted a series of magnetohydrodynamic simulations of CMEs on a solar-type star of moderate activity levels. The propagation and evolution of the CMEs were traced in the three-dimensional outputs and the temporal evolution of their dynamic properties (such as masses, velocities, and kinetic energies) were determined. The simulated stellar CMEs are more massive and energetic than their solar analog, which is a result of the stronger magnetic field on the surface of the simulated star than that of the Sun. The simulated CMEs display masses ranging from ~10^16 g to ~10^18 g and kinetic energies from ~10^31 erg to ~10^33 erg. We also investigated the instantaneous influence of the CMEs to the star's angular momentum loss rate. Our results suggest that angular momentum can either be added to or be removed from the star during the evolution of CME events. We found a positive correlation between the amplitude of the angular momentum loss rate variation and the CME's kinetic energy as well as mass, suggesting that more energetic/massive CMEs have higher possibility to add angular momentum to the star.

astro-ph.SR↗

arXiv2024

Sun-as-a-star observations of obscuration dimmings caused by filament eruptions

Filament eruptions often lead to coronal mass ejections (CMEs) on the Sun and are one of the most energetic eruptive phenomena in the atmospheres of other late-type stars. However, the detection of filament eruptions and CMEs on stars beyond the solar system is challenging. Here we present six filament eruption cases on the Sun and show that filament material obscuring part of the solar disk can cause detectable dimming signatures in sun-as-a-star flux curves of He II 304 A. Those filament eruptions have similar morphological features, originating from small filaments inside active regions and subsequently strongly expanding to obscure large areas of the solar disk or the bright flare regions. We have tracked the detailed evolution of six obscuration dimmings and estimated the dimming properties, such as dimming depths, dimming areas, and duration. The largest dimming depth among the six events under study is 6.2% accompanied by the largest dimming area of 5.6\% of the solar disk area. Other events have maximum dimming depths in a range of around 1% to 3% with maximum areas varying between about 3% to 4% of the solar disk area. The duration of the dimming spans from around 0.4 hours to 7.0 hours for the six events under study. A positive correlation was found between the dimming depth and area, which may help to set constraint on the filament sizes in stellar observations.

astro-ph.SR↗

arXiv2024

A Local Gaussian Process Regression Approach to Frequency Response Function Estimation

Frequency response function (FRF) estimation is a classical subject in system identification. In the past two decades, there have been remarkable advances in developing local methods for this subject, e.g., the local polynomial method, local rational method, and iterative local rational method. The recent concentrations for local methods are two issues: the model order selection and the identification of lightly damped systems. To address these two issues, we propose a new local method called local Gaussian process regression (LGPR). We show that the frequency response function locally is either analytic or resonant, and this prior knowledge can be embedded into a kernel-based regularized estimate through a dot-product kernel plus a resonance kernel induced by a second-order resonant system. The LGPR provides a new route to tackle the aforementioned issues. In the numerical simulations, the LGPR shows the best FRF estimation accuracy compared with the existing local methods, and moreover, the LGPR is more robust with respect to sample size and noise level.

eess.SY↗

arXiv2024

Transnational Network Dynamics of Problematic Information Diffusion

This study maps the spread of two cases of COVID-19 conspiracy theories and misinformation in Spanish and French in Latin American and French-speaking communities on Facebook, and thus contributes to understanding the dynamics, reach and consequences of emerging transnational misinformation networks. The findings show that co-sharing behavior of public Facebook groups created transnational networks by sharing videos of Medicos por la Verdad (MPV) conspiracy theories in Spanish and hydroxychloroquine-related misinformation sparked by microbiologist Didier Raoult (DR) in French, usually igniting the surge of locally led interest groups across the Global South. Using inferential methods, the study shows how these networks are enabled primarily by shared cultural and thematic attributes among Facebook groups, effectively creating very large, networked audiences. The study contributes to the understanding of how potentially harmful conspiracy theories and misinformation transcend national borders through non-English speaking online communities, further highlighting the overlooked role of transnationalism in global misinformation diffusion and the potentially disproportionate harm that it causes in vulnerable communities across the globe.

cs.SI↗

arXiv2024

Break-for-Make: Modular Low-Rank Adaptations for Composable Content-Style Customization

Personalized generation paradigms empower designers to customize visual intellectual properties with the help of textual descriptions by tuning or adapting pre-trained text-to-image models on a few images. Recent works explore approaches for concurrently customizing both content and detailed visual style appearance. However, these existing approaches often generate images where the content and style are entangled. In this study, we reconsider the customization of content and style concepts from the perspective of parameter space construction. Unlike existing methods that utilize a shared parameter space for content and style, we propose a learning framework that separates the parameter space to facilitate individual learning of content and style, thereby enabling disentangled content and style. To achieve this goal, we introduce "partly learnable projection" (PLP) matrices to separate the original adapters into divided sub-parameter spaces. We propose "break-for-make" customization learning pipeline based on PLP, which is simple yet effective. We break the original adapters into "up projection" and "down projection", train content and style PLPs individually with the guidance of corresponding textual prompts in the separate adapters, and maintain generalization by employing a multi-correspondence projection learning strategy. Based on the adapters broken apart for separate training content and style, we then make the entity parameter space by reconstructing the content and style PLPs matrices, followed by fine-tuning the combined adapter to generate the target object with the desired appearance. Experiments on various styles, including textures, materials, and artistic style, show that our method outperforms state-of-the-art single/multiple concept learning pipelines in terms of content-style-prompt alignment.

cs.CV↗

arXiv2024

An ultraviolet photodetector based on conductive hydrogenated TiO$_2$ film prepared by radio frequency atmospheric pressure plasma

The growing demand for real-time ultraviolet (UV) monitoring calls for a simple, rapid, and low-cost strategy to prepare UV photodetectors. We prepare a wearable real-time UV photodetector based on hydrogenated titanium dioxide film synthesized by radio frequency atmospheric pressure plasma. The conductivity of our hydrogenated titanium dioxide is improved to 10.2 S cm$^{-1}$, 9 orders of magnitude higher than that of pristine titanium dioxide after 10 min plasma treatment. Plasma hydrogenation disrupts the surface crystal structure, introducing oxygen vacancies (OVs) that create self-doped titanium(III) and titanium (II) species. First-principles calculations indicate the OVs raise the Fermi level of TiO$_2$ and distort the lattice nearby. Our optimized film has a distinctive periodic switching characteristic under intermittent illumination and good responsivity from 280 - 400 nm, peaking at 632.35 mA W^{-1}$ at 365 nm. The fabricated wearable sensor based on the optimized film effectively performs the monitoring of the daily variation of ambient UV intensity in three typical weather types and transferring its data to a smartphone via Wi-Fi.

physics.plasm-ph↗

arXiv2024

SIFiD: Reassess Summary Factual Inconsistency Detection with LLM

Ensuring factual consistency between the summary and the original document is paramount in summarization tasks. Consequently, considerable effort has been dedicated to detecting inconsistencies. With the advent of Large Language Models (LLMs), recent studies have begun to leverage their advanced language understanding capabilities for inconsistency detection. However, early attempts have shown that LLMs underperform traditional models due to their limited ability to follow instructions and the absence of an effective detection methodology. In this study, we reassess summary inconsistency detection with LLMs, comparing the performances of GPT-3.5 and GPT-4. To advance research in LLM-based inconsistency detection, we propose SIFiD (Summary Inconsistency Detection with Filtered Document) that identify key sentences within documents by either employing natural language inference or measuring semantic similarity between summaries and documents.

cs.CL↗

arXiv2024

Is it possible to detect coronal mass ejections on solar-type stars through extreme-ultraviolet spectral observations?

Stellar coronal mass ejections (CMEs) from host stars are an important factor that affects the habitability of exoplanets. Although their solar counterparts have been well observed for decades, it is still very difficult to find solid evidence of stellar CMEs. Using the spectral line profile asymmetry caused by the Doppler shift of erupting plasma, several stellar CME candidates have been identified from spectral lines formed at chromospheric or transition region temperatures of the stars. However, a successful detection of stellar CME signals based on the profile asymmetries of coronal lines is still lacking. It is unclear whether we can detect such signals. Here we construct an analytical model for CMEs on solar-type stars, and derive an expression of stellar extreme-ultraviolet (EUV) line profiles during CME eruptions. For different instrumental parameters, exposure times, CME propagation directions and stellar activity levels, we synthesized the corresponding line profiles of Fe IX 171.07 Å and Fe XV 284.16 Å. Further investigations provide constraints on the instrumental requirements for successful detection and characterization of stellar CMEs. Our results show that it is possible to detect stellar CME signals and infer their velocities based on spectral profile asymmetries using an EUV spectrometer with a moderate spectral resolution and signal-to-noise ratio. Our work provides important references for the design of future EUV spectrometers for stellar CME detection and the development of observation strategies.

astro-ph.SR↗

arXiv2024

A ThermalKinetic Inductance Detectors Pixel Design for Cosmic Microwave Background Observations at 90/150 GHz bands

The highly sensitive millimeter-wave telescope is an important tool for accurate measurement of Cosmic Microwave Background (CMB) radiation, and its core component is a detector array located in a cryogenic focal plane. The feasibility of utilizing thermal kinetic inductance detectors (TKIDs) for CMB observations has been demonstrated. We propose a pixel design of TKIDs for observing CMB through atmospheric windows for observations in the 90/150 GHz bands. Assuming lossless dielectric, the coupling efficiency of a single pixel is around 90%. This pixel design will be utilized for future large-scale TKIDs array designs for CMB observations.

astro-ph.IM↗

arXiv2023

UniGS: Unified Representation for Image Generation and Segmentation

This paper introduces a novel unified representation of diffusion models for image generation and segmentation. Specifically, we use a colormap to represent entity-level masks, addressing the challenge of varying entity numbers while aligning the representation closely with the image RGB domain. Two novel modules, including the location-aware color palette and progressive dichotomy module, are proposed to support our mask representation. On the one hand, a location-aware palette guarantees the colors' consistency to entities' locations. On the other hand, the progressive dichotomy module can efficiently decode the synthesized colormap to high-quality entity-level masks in a depth-first binary search without knowing the cluster numbers. To tackle the issue of lacking large-scale segmentation training data, we employ an inpainting pipeline and then improve the flexibility of diffusion models across various tasks, including inpainting, image synthesis, referring segmentation, and entity segmentation. Comprehensive experiments validate the efficiency of our approach, demonstrating comparable segmentation mask quality to state-of-the-art and adaptability to multiple tasks. The code will be released at \href{https://github.com/qqlu/Entity}{https://github.com/qqlu/Entity}.

cs.CV↗

arXiv2023

GOAL: A Challenging Knowledge-grounded Video Captioning Benchmark for Real-time Soccer Commentary Generation

Despite the recent emergence of video captioning models, how to generate vivid, fine-grained video descriptions based on the background knowledge (i.e., long and informative commentary about the domain-specific scenes with appropriate reasoning) is still far from being solved, which however has great applications such as automatic sports narrative. In this paper, we present GOAL, a benchmark of over 8.9k soccer video clips, 22k sentences, and 42k knowledge triples for proposing a challenging new task setting as Knowledge-grounded Video Captioning (KGVC). Moreover, we conduct experimental adaption of existing methods to show the difficulty and potential directions for solving this valuable and applicable task. Our data and code are available at https://github.com/THU-KEG/goal.

cs.CV↗

arXiv2023

Development of a SciFi-based beam monitor for COMET

COMET is a leading experiment to search for coherent conversion of $μ^- \mathrm{N}\to e^- \mathrm{N}$ with a high-intensity pulsed muon beamline, produced by the innovative slow extraction techniques. Therefore, it is critical to measure the characteristics of the muon beam. We set up a Muon Beam Monitor (MBM), where scintillation fibers (SciFi) weaved in the cross shape are coupled to silicon photomultipliers (SiPM), to measure the spatial profile and timing structure of the extracted muon beam for COMET. The MBM detector has been tested successfully with a proton beamline in China Spallation Neutron Source (CSNS) and taken data with good performance in the commissioning run called COMET Phase-$α$. Experience of the MBM development, such as the mechanical structure and electronics readout, and its beam measurement results will be shared.

physics.ins-det↗

arXiv2023

Tenrec: A Large-scale Multipurpose Benchmark Dataset for Recommender Systems

Existing benchmark datasets for recommender systems (RS) either are created at a small scale or involve very limited forms of user feedback. RS models evaluated on such datasets often lack practical values for large-scale real-world applications. In this paper, we describe Tenrec, a novel and publicly available data collection for RS that records various user feedback from four different recommendation scenarios. To be specific, Tenrec has the following five characteristics: (1) it is large-scale, containing around 5 million users and 140 million interactions; (2) it has not only positive user feedback, but also true negative feedback (vs. one-class recommendation); (3) it contains overlapped users and items across four different scenarios; (4) it contains various types of user positive feedback, in forms of clicks, likes, shares, and follows, etc; (5) it contains additional features beyond the user IDs and item IDs. We verify Tenrec on ten diverse recommendation tasks by running several classical baseline models per task. Tenrec has the potential to become a useful benchmark dataset for a majority of popular recommendation tasks.

cs.IR↗