arXiv subjects

Yuqian Liu

Publications and source records attributed to Yuqian Liu.

9 recordsLinked to original sources

arXiv2025

Soft Coulomb Gap Limits the Performance of Organic Thermoelectrics

Although consensus exists that the thermoelectric properties of doped organic semiconductors result from a complex interplay between a large number of mutually dependent factors, there is no consensus on which of these are dominant, or even on how to best describe the charge and energy transport. This holds particularly in the intermediate doping regime where the optimal performance is typically observed at the roll-off in the Seebeck coefficient - conductivity (S-{\sigma} correlation, fundamentally limiting the rational advancement of organic thermoelectric materials. Here, we combine experiments on a board set of conjugated polymers with kinetic Monte Carlo simulations across varying doping levels to uncover a general transport framework. We demonstrate that the optimal thermoelectric power factor (PF_max) consistently occurs at the transition between conventional variable-range hopping (VRH) and VRH in a density of states in which a soft Coulomb gap forms at the Fermi level, as described by Efros and Shklovskii (ES-VRH). This suggests the use of high dielectric constant materials or the promotion of charge delocalization as an avenue to shift the roll-off of the S-{\sigma}} curve, which constrains PF_max, to higher doping levels and accordingly higher PF.

cond-mat.mtrl-sci↗

arXiv2025

Exploring the Hidden Reasoning Process of Large Language Models by Misleading Them

Large language models (LLMs) have been able to perform various forms of reasoning tasks in a wide range of scenarios, but are they truly engaging in task abstraction and rule-based reasoning beyond mere memorization? To answer this question, we propose a novel experimental approach, Misleading Fine-Tuning (MisFT), to examine whether LLMs perform abstract reasoning by altering their original understanding of fundamental rules. In particular, by constructing datasets with math expressions or logical formulas that contradict correct principles, we fine-tune the model to learn those contradictory rules and assess its generalization ability on unseen test domains. Through a series of experiments, we find that current LLMs are capable of applying contradictory rules to solve practical math word problems and natural language reasoning tasks, implying the presence of an internal mechanism in LLMs that abstracts before reasoning.

cs.LG↗

arXiv2023

Vox-Fusion: Dense Tracking and Mapping with Voxel-based Neural Implicit Representation

In this work, we present a dense tracking and mapping system named Vox-Fusion, which seamlessly fuses neural implicit representations with traditional volumetric fusion methods. Our approach is inspired by the recently developed implicit mapping and positioning system and further extends the idea so that it can be freely applied to practical scenarios. Specifically, we leverage a voxel-based neural implicit surface representation to encode and optimize the scene inside each voxel. Furthermore, we adopt an octree-based structure to divide the scene and support dynamic expansion, enabling our system to track and map arbitrary scenes without knowing the environment like in previous works. Moreover, we proposed a high-performance multi-process framework to speed up the method, thus supporting some applications that require real-time performance. The evaluation results show that our methods can achieve better accuracy and completeness than previous methods. We also show that our Vox-Fusion can be used in augmented reality and virtual reality applications. Our source code is publicly available at https://github.com/zju3dv/Vox-Fusion.

cs.CV↗

arXiv2023

Vox-Surf: Voxel-based Implicit Surface Representation

Virtual content creation and interaction play an important role in modern 3D applications such as AR and VR. Recovering detailed 3D models from real scenes can significantly expand the scope of its applications and has been studied for decades in the computer vision and computer graphics community. We propose Vox-Surf, a voxel-based implicit surface representation. Our Vox-Surf divides the space into finite bounded voxels. Each voxel stores geometry and appearance information in its corner vertices. Vox-Surf is suitable for almost any scenario thanks to sparsity inherited from voxel representation and can be easily trained from multiple view images. We leverage the progressive training procedure to extract important voxels gradually for further optimization so that only valid voxels are preserved, which greatly reduces the number of sampling points and increases rendering speed.The fine voxels can also be considered as the bounding volume for collision detection.The experiments show that Vox-Surf representation can learn delicate surface details and accurate color with less memory and faster rendering speed than other methods.We also show that Vox-Surf can be more practical in scene editing and AR applications.

cs.CV↗

arXiv2022

OpenCalib: A Multi-sensor Calibration Toolbox for Autonomous Driving

Accurate sensor calibration is a prerequisite for multi-sensor perception and localization systems for autonomous vehicles. The intrinsic parameter calibration of the sensor is to obtain the mapping relationship inside the sensor, and the extrinsic parameter calibration is to transform two or more sensors into a unified spatial coordinate system. Most sensors need to be calibrated after installation to ensure the accuracy of sensor measurements. To this end, we present OpenCalib, a calibration toolbox that contains a rich set of various sensor calibration methods. OpenCalib covers manual calibration tools, automatic calibration tools, factory calibration tools, and online calibration tools for different application scenarios. At the same time, to evaluate the calibration accuracy and subsequently improve the accuracy of the calibration algorithm, we released a corresponding benchmark dataset. This paper introduces various features and calibration methods of this toolbox. To our knowledge, this is the first open-sourced calibration codebase containing the full set of autonomous-driving-related calibration approaches in this area. We wish that the toolbox could be helpful to autonomous driving researchers. We have open-sourced our code on GitHub to benefit the community. Code is available at https://github.com/PJLab-ADG/SensorsCalibration.

cs.RO↗

arXiv2017

A GRB and Broad-lined Type Ic Supernova from a Single Central Engine

Unusually high velocities (< ~0.1c) and correspondingly high kinetic energies have been observed in a subset of Type Ic supernovae (so-called "broad-lined Ic" supernovae; SNe Ic-BL), prompting a search for a central engine model capable of generating such energetic explosions. A clue to the explosion mechanism may lie in the fact that all supernovae that accompany long-duration gamma-ray bursts belong to the SN Ic-BL class. Using a combination of two-dimensional relativistic hydrodynamics and radiation transport calculations, we demonstrate that the central engine responsible for long gamma-ray bursts can also trigger a SN Ic-BL. We find that a reasonable gamma-ray burst engine injected into a stripped Wolf-Rayet progenitor produces a relativistic jet with energy ~10^51 ergs, as well as a SN whose synthetic light curves and spectra are fully consistent with observed SNe Ic-BL during the photospheric phase. As a result of the jet's asymmetric energy injection, the SN spectra and light curves depend on viewing angle. The impact of viewing angle on the spectrum is particularly pronounced at early times, while the viewing angle dependence for the light curves (~10% variation in bolometric luminosity) persists throughout the photospheric phase.

astro-ph.HE↗

arXiv2017

Revisiting the Lick Observatory Supernova Search Volume-Limited Sample: Updated Classifications and Revised Stripped-Envelope Supernova Fractions

We re-examine the classifications of supernovae (SNe) presented in the Lick Observatory Supernova Search (LOSS) volume-limited sample with a focus on the stripped-envelope SNe. The LOSS volume-limited sample, presented by Leaman et al. (2011) and Li et al. (2011b), was calibrated to provide meaningful measurements of SN rates in the local universe; the results presented therein continue to be used for comparisons to theoretical and modeling efforts. Many of the objects from the LOSS sample were originally classified based upon only a small subset of the data now available, however, and recent studies have both updated some subtype distinctions and improved our ability to perform robust classications, especially for stripped-envelope SNe. We re-examine the spectroscopic classifications of all events in the LOSS volume-limited sample (180 SNe and SN impostors) and update them if necessary. We discuss the populations of rare objects in our sample including broad-lined Type Ic SNe, Ca-rich SNe, SN 1987A-like events (we identify SN 2005io as SN 1987A-like here for the first time), and peculiar subtypes. The relative fractions of Type Ia SNe, Type II SNe, and stripped-envelope SNe in the local universe are not affected, but those of some subtypes are. Most significantly, after discussing the often unclear boundary between SNe Ib and Ic when only noisy spectra are available, we find a higher SN Ib fraction and a lower SN Ic fraction than calculated by Li et al. (2011b): spectroscopically normal SNe Ib occur in the local universe $1.7 \pm 0.9$ times more often than do normal SNe Ic.

astro-ph.HE↗

arXiv2016

Monte Carlo Method for Calculating Oxygen Abundances and Their Uncertainties from Strong-Line Flux Measurements

We present the open-source Python code pyMCZ that determines oxygen abundance and its distribution from strong emission lines in the standard metallicity calibrators, based on the original IDL code of Kewley & Dopita (2002) with updates from Kewley & Ellison (2008), and expanded to include more recently developed calibrators. The standard strong-line diagnostics have been used to estimate the oxygen abundance in the interstellar medium through various emission line ratios in many areas of astrophysics, including galaxy evolution and supernova host galaxy studies. We introduce a Python implementation of these methods that, through Monte Carlo sampling, better characterizes the statistical oxygen abundance confidence region including the effect due to the propagation of observational uncertainties. These uncertainties are likely to dominate the error budget in the case of distant galaxies, hosts of cosmic explosions. Given line flux measurements and their uncertainties, our code produces synthetic distributions for the oxygen abundance in up to 15 metallicity calibrators simultaneously, as well as for E(B-V), and estimates their median values and their 68% confidence regions. We test our code on emission line measurements from a sample of nearby supernova host galaxies (z < 0.15) and compare our metallicity results with those from previous methods. Our metallicity estimates are consistent with previous methods but yield smaller statistical uncertainties. Systematic uncertainties are not taken into account. We offer visualization tools to assess the spread of the oxygen abundance in the different calibrators, as well as the shape of the estimated oxygen abundance distribution in each calibrator, and develop robust metrics for determining the appropriate Monte Carlo sample size. The code is open access and open source and can be found at https://github.com/nyusngroup/pyMCZ (Abridged)

astro-ph.IM↗

arXiv2015

SuperNova IDentification spectral templates of 70 stripped-envelope core-collapse supernovae

We constructed 70 SuperNova IDentification (SNID; Blondin & Tonry 2007) supernova (SN) templates using 640 spectra of stripped-envelope core-collapse SNe (SESNe) published by Modjaz et al. (2014). Fifty-six SN templates which are constructed from 458 spectra are brand new, increasing the number of SESNe and the number of SESNe spectra in the current SNID database by a factor of 2.6 and 2.0, respectively. We also made some type and phase refinements to templates in the current SNID database.

astro-ph.SR↗