arXiv subjects

Shuo Sun

Publications and source records attributed to Shuo Sun.

At least 109 records · Page 6Linked to original sources

arXiv2023

Nitrogen-vacancy magnetometry of individual Fe-triazole spin crossover nanorods

[Fe(Htrz)2(trz)](BF4) (Fe-triazole) spin crossover molecules show thermal, electrical, and optical switching between high spin (HS) and low spin (LS) states, making them promising candidates for molecular spintronics. The LS and HS transitions originate from the electronic configurations of Fe(II), and are considered to be diamagnetic and paramagnetic respectively. The Fe(II) LS state has six paired electrons in the ground states with no interaction with the magnetic field and a diamagnetic behavior is usually observed. While the bulk magnetic properties of Fe-triazole compounds are widely studied by standard magnetometry techniques their properties at the individual level are missing. Here we use nitrogen vacancy (NV) based magnetometry to study the magnetic properties of the Fe-triazole LS state of nanoparticle clusters and individual nanorods of size varying from 20 to 1000 nm. Scanning electron microscopy (SEM) and Raman spectroscopy are performed to determine the size of the nanoparticles/nanorods and to confirm their respective spin state. The magnetic field patterns produced by the nanoparticles/nanorods are imaged by NV magnetic microscopy as a function of applied magnetic field (up to 350 mT) and correlated with SEM and Raman. We found that in most of the nanorods the LS state is slightly paramagnetic, possibly originating from the surface oxidation and/or the greater Fe(III) presence along the nanorod edges. NV measurements on the Fe-triazole LS state nanoparticle clusters revealed both diamagnetic and paramagnetic behavior. Our results highlight the potential of NV quantum sensors to study the magnetic properties of spin crossover molecules and molecular magnets.

cond-mat.mes-hall↗

arXiv2023

PrefRec: Recommender Systems with Human Preferences for Reinforcing Long-term User Engagement

Current advances in recommender systems have been remarkably successful in optimizing immediate engagement. However, long-term user engagement, a more desirable performance metric, remains difficult to improve. Meanwhile, recent reinforcement learning (RL) algorithms have shown their effectiveness in a variety of long-term goal optimization tasks. For this reason, RL is widely considered as a promising framework for optimizing long-term user engagement in recommendation. Though promising, the application of RL heavily relies on well-designed rewards, but designing rewards related to long-term user engagement is quite difficult. To mitigate the problem, we propose a novel paradigm, recommender systems with human preferences (or Preference-based Recommender systems), which allows RL recommender systems to learn from preferences about users historical behaviors rather than explicitly defined rewards. Such preferences are easily accessible through techniques such as crowdsourcing, as they do not require any expert knowledge. With PrefRec, we can fully exploit the advantages of RL in optimizing long-term goals, while avoiding complex reward engineering. PrefRec uses the preferences to automatically train a reward function in an end-to-end manner. The reward function is then used to generate learning signals to train the recommendation policy. Furthermore, we design an effective optimization method for PrefRec, which uses an additional value function, expectile regression and reward model pre-training to improve the performance. We conduct experiments on a variety of long-term user engagement optimization tasks. The results show that PrefRec significantly outperforms previous state-of-the-art methods in all the tasks.

cs.IR↗

arXiv2023

PRUDEX-Compass: Towards Systematic Evaluation of Reinforcement Learning in Financial Markets

The financial markets, which involve more than $90 trillion market capitals, attract the attention of innumerable investors around the world. Recently, reinforcement learning in financial markets (FinRL) has emerged as a promising direction to train agents for making profitable investment decisions. However, the evaluation of most FinRL methods only focuses on profit-related measures and ignores many critical axes, which are far from satisfactory for financial practitioners to deploy these methods into real-world financial markets. Therefore, we introduce PRUDEX-Compass, which has 6 axes, i.e., Profitability, Risk-control, Universality, Diversity, rEliability, and eXplainability, with a total of 17 measures for a systematic evaluation. Specifically, i) we propose AlphaMix+ as a strong FinRL baseline, which leverages mixture-of-experts (MoE) and risk-sensitive approaches to make diversified risk-aware investment decisions, ii) we evaluate 8 FinRL methods in 4 long-term real-world datasets of influential financial markets to demonstrate the usage of our PRUDEX-Compass, iii) PRUDEX-Compass together with 4 real-world datasets, standard implementation of 8 FinRL methods and a portfolio management environment is released as public resources to facilitate the design and comparison of new FinRL methods. We hope that PRUDEX-Compass can not only shed light on future FinRL research to prevent untrustworthy results from stagnating FinRL into successful industry deployment but also provide a new challenging algorithm evaluation scenario for the reinforcement learning (RL) community.

q-fin.TR↗

arXiv2023

Detecting the gravitational wave memory effect with TianQin

The gravitational wave memory effect is a prediction of general relativity. The presence of memory effect in gravitational wave signals not only provides the chance to test an important aspect of general relativity, but also represents a potentially non-negligible contribution to the waveform for certain gravitational wave events. In this paper, we study the prospect of detecting the gravitational wave memory effect directly with the planned space-based gravitational wave detector -- TianQin. We find that during its 5 years operation, for the gravitational wave signals that could be detected by TianQin, about $0.5\sim2.0$ signals may contain displacement memory effect with signal-to-noise ratios (SNRs) greater than 3. This suggests that the chance for TianQin to detect the displacement memory effect directly is low but not fully negligible. In contrast, the chance to detect the spin memory is negligible. We also study that in which parameter space, the memory effect is expected to be significant in waveform modeling.

gr-qc↗

arXiv2023

Performance analysis of quantum repeaters enabled by deterministically generated photonic graph states

By encoding logical qubits into specific types of photonic graph states, one can realize quantum repeaters that enable fast entanglement distribution rates approaching classical communication. However, the generation of these photonic graph states requires a formidable resource overhead using traditional approaches based on linear optics. Overcoming this challenge, a number of new schemes have been proposed that employ quantum emitters to deterministically generate photonic graph states. Although these schemes have the potential to significantly reduce the resource cost, a systematic comparison of the repeater performance among different encodings and different generation schemes is lacking. Here, we quantitatively analyze the performance of quantum repeaters based on two different graph states, i.e. the tree graph states and the repeater graph states. For both states, we compare the performance between two generation schemes, one based on a single quantum emitter coupled to ancillary matter qubits, and one based on a single quantum emitter coupled to a delayed feedback. We identify the numerically optimal scheme at different system parameters. Our analysis provides a clear guideline on the selection of the generation scheme for graph-state-based quantum repeaters, and lays out the parameter requirements for future experimental realizations of different schemes.

quant-ph↗

arXiv2022

GET-DIPP: Graph-Embedded Transformer for Differentiable Integrated Prediction and Planning

Accurately predicting interactive road agents' future trajectories and planning a socially compliant and human-like trajectory accordingly are important for autonomous vehicles. In this paper, we propose a planning-centric prediction neural network, which takes surrounding agents' historical states and map context information as input, and outputs the joint multi-modal prediction trajectories for surrounding agents, as well as a sequence of control commands for the ego vehicle by imitation learning. An agent-agent interaction module along the time axis is proposed in our network architecture to better comprehend the relationship among all the other intelligent agents on the road. To incorporate the map's topological information, a Dynamic Graph Convolutional Neural Network (DGCNN) is employed to process the road network topology. Besides, the whole architecture can serve as a backbone for the Differentiable Integrated motion Prediction with Planning (DIPP) method by providing accurate prediction results and initial planning commands. Experiments are conducted on real-world datasets to demonstrate the improvements made by our proposed method in both planning and prediction accuracy compared to the previous state-of-the-art methods.

cs.RO↗

arXiv2022

DeepScalper: A Risk-Aware Reinforcement Learning Framework to Capture Fleeting Intraday Trading Opportunities

Reinforcement learning (RL) techniques have shown great success in many challenging quantitative trading tasks, such as portfolio management and algorithmic trading. Especially, intraday trading is one of the most profitable and risky tasks because of the intraday behaviors of the financial market that reflect billions of rapidly fluctuating capitals. However, a vast majority of existing RL methods focus on the relatively low frequency trading scenarios (e.g., day-level) and fail to capture the fleeting intraday investment opportunities due to two major challenges: 1) how to effectively train profitable RL agents for intraday investment decision-making, which involves high-dimensional fine-grained action space; 2) how to learn meaningful multi-modality market representation to understand the intraday behaviors of the financial market at tick-level. Motivated by the efficient workflow of professional human intraday traders, we propose DeepScalper, a deep reinforcement learning framework for intraday trading to tackle the above challenges. Specifically, DeepScalper includes four components: 1) a dueling Q-network with action branching to deal with the large action space of intraday trading for efficient RL optimization; 2) a novel reward function with a hindsight bonus to encourage RL agents making trading decisions with a long-term horizon of the entire trading day; 3) an encoder-decoder architecture to learn multi-modality temporal market embedding, which incorporates both macro-level and micro-level market information; 4) a risk-aware auxiliary task to maintain a striking balance between maximizing profit and minimizing risk. Through extensive experiments on real-world market data spanning over three years on six financial futures, we demonstrate that DeepScalper significantly outperforms many state-of-the-art baselines in terms of four financial criteria.

q-fin.TR↗

arXiv2022

Quantitative Stock Investment by Routing Uncertainty-Aware Trading Experts: A Multi-Task Learning Approach

Quantitative investment is a fundamental financial task that highly relies on accurate stock prediction and profitable investment decision making. Despite recent advances in deep learning (DL) have shown stellar performance on capturing trading opportunities in the stochastic stock market, we observe that the performance of existing DL methods is sensitive to random seeds and network initialization. To design more profitable DL methods, we analyze this phenomenon and find two major limitations of existing works. First, there is a noticeable gap between accurate financial predictions and profitable investment strategies. Second, investment decisions are made based on only one individual predictor without consideration of model uncertainty, which is inconsistent with the workflow in real-world trading firms. To tackle these two limitations, we first reformulate quantitative investment as a multi-task learning problem. Later on, we propose AlphaMix, a novel two-stage mixture-of-experts (MoE) framework for quantitative investment to mimic the efficient bottom-up trading strategy design workflow of successful trading firms. In Stage one, multiple independent trading experts are jointly optimized with an individual uncertainty-aware loss function. In Stage two, we train neural routers (corresponding to the role of a portfolio manager) to dynamically deploy these experts on an as-needed basis. AlphaMix is also a universal framework that is applicable to various backbone network architectures with consistent performance gains. Through extensive experiments on long-term real-world data spanning over five years on two of the most influential financial markets (US and China), we demonstrate that AlphaMix significantly outperforms many state-of-the-art baselines in terms of four financial criteria.

q-fin.TR↗

arXiv2022

Value Functions for Depth-Limited Solving in Zero-Sum Imperfect-Information Games

We provide a formal definition of depth-limited games together with an accessible and rigorous explanation of the underlying concepts, both of which were previously missing in imperfect-information games. The definition works for an arbitrary extensive-form game and is not tied to any specific game-solving algorithm. Moreover, this framework unifies and significantly extends three approaches to depth-limited solving that previously existed in extensive-form games and multiagent reinforcement learning but were not known to be compatible. A key ingredient of these depth-limited games are value functions. Focusing on two-player zero-sum imperfect-information games, we show how to obtain optimal value functions and prove that public information provides both necessary and sufficient context for computing them. We provide a domain-independent encoding of the domains that allows for approximating value functions even by simple feed-forward neural networks, which are then able to generalize to unseen parts of the game. We use the resulting value network to implement a depth-limited version of counterfactual regret minimization. In three distinct domains, we show that the algorithm's exploitability is roughly linearly dependent on the value network's quality and that it is not difficult to train a value network with which depth-limited CFR's performance is as good as that of CFR with access to the full game.

cs.AI↗

arXiv2021

Epitaxial Growth of Ultraflat Bismuthene with Large Topological Band Inversion Enabled by Substrate-Orbital-Filtering Effect

Quantum spin Hall (QSH) systems hold promises of low-power-consuming spintronic devices, yet their practical applications are extremely impeded by the small energy gaps. Fabricating QSH materials with large gaps, especially under the guidance of design principles, is essential for both scientific research and practical applications. Here, we demonstrate that large on-site atomic spin-orbit coupling can be directly exploited via the intriguing substrate-orbital-filtering effect to generate large-gap QSH systems and experimentally realized on the epitaxially synthesized ultraflat bismuthene on Ag(111). Theoretical calculations reveal that the underlying substrate selectively filters Bi pz orbitals away from the Fermi level, leading pxy orbitals with nonzero magnetic quantum numbers, resulting in large topological gap of ~1 eV at the K point. The corresponding topological edge states are identified through scanning tunneling spectroscopy combined with density functional theory calculations. Our findings provide general strategies to design large-gap QSH systems and further explore their topology-related physics.

cond-mat.mes-hall↗

arXiv2021

MLQE-PE: A Multilingual Quality Estimation and Post-Editing Dataset

We present MLQE-PE, a new dataset for Machine Translation (MT) Quality Estimation (QE) and Automatic Post-Editing (APE). The dataset contains eleven language pairs, with human labels for up to 10,000 translations per language pair in the following formats: sentence-level direct assessments and post-editing effort, and word-level good/bad labels. It also contains the post-edited sentences, as well as titles of the articles where the sentences were extracted from, and the neural MT models used to translate the text.

cs.CL↗

arXiv2021

Reinforcement Learning for Quantitative Trading

Quantitative trading (QT), which refers to the usage of mathematical models and data-driven techniques in analyzing the financial market, has been a popular topic in both academia and financial industry since 1970s. In the last decade, reinforcement learning (RL) has garnered significant interest in many domains such as robotics and video games, owing to its outstanding ability on solving complex sequential decision making problems. RL's impact is pervasive, recently demonstrating its ability to conquer many challenging QT tasks. It is a flourishing research direction to explore RL techniques' potential on QT tasks. This paper aims at providing a comprehensive survey of research efforts on RL-based methods for QT tasks. More concretely, we devise a taxonomy of RL-based QT models, along with a comprehensive summary of the state of the art. Finally, we discuss current challenges and propose future research directions in this exciting field.

cs.LG↗

arXiv2021

Giant Transport Anisotropy in ReS$_2$ Revealed via Nanoscale Conducting Path Control

The low in-plane symmetry in layered 1T'-ReS$_2$ results in strong band anisotropy, while its manifestation in the electronic properties is challenging to resolve due to the lack of effective approaches for controlling the local current path. In this work, we reveal the giant transport anisotropy in monolayer to four-layer ReS$_2$ by creating directional conducting paths via nanoscale ferroelectric control. By reversing the polarization of a ferroelectric polymer top layer, we induce conductivity switching ratio of >1.5x10$^8$ in the ReS$_2$ channel at 300 K. Characterizing the domain-defined conducting nanowires in an insulating background shows that the conductivity ratio between the directions along and perpendicular to the Re-chain can exceed 5.5x10$^4$. Theoretical modeling points to the band origin of the transport anomaly, and further reveals the emergence of a flat band in few-layer ReS$_2$. Our work paves the path for implementing the highly anisotropic 2D materials for designing novel collective phenomena and electron lensing applications.

cond-mat.mes-hall↗

arXiv2021

Classification-based Quality Estimation: Small and Efficient Models for Real-world Applications

Sentence-level Quality estimation (QE) of machine translation is traditionally formulated as a regression task, and the performance of QE models is typically measured by Pearson correlation with human labels. Recent QE models have achieved previously-unseen levels of correlation with human judgments, but they rely on large multilingual contextualized language models that are computationally expensive and make them infeasible for real-world applications. In this work, we evaluate several model compression techniques for QE and find that, despite their popularity in other NLP tasks, they lead to poor performance in this regression setting. We observe that a full model parameterization is required to achieve SoTA results in a regression task. However, we argue that the level of expressiveness of a model in a continuous range is unnecessary given the downstream applications of QE, and show that reframing QE as a classification problem and evaluating QE models using classification metrics would better reflect their actual performance in real-world applications.

cs.CL↗

arXiv2021

Causal Inference for Quantile Treatment Effects

Analyses of environmental phenomena often are concerned with understanding unlikely events such as floods, heatwaves, droughts or high concentrations of pollutants. Yet the majority of the causal inference literature has focused on modelling means, rather than (possibly high) quantiles. We define a general estimator of the population quantile treatment (or exposure) effects (QTE) -- the weighted QTE (WQTE) -- of which the population QTE is a special case, along with a general class of balancing weights incorporating the propensity score. Asymptotic properties of the proposed WQTE estimators are derived. We further propose and compare propensity score regression and two weighted methods based on these balancing weights to understand the causal effect of an exposure on quantiles, allowing for the exposure to be binary, discrete or continuous. Finite sample behavior of the three estimators is studied in simulation. The proposed methods are applied to data taken from the Bavarian Danube catchment area to estimate the 95% QTE of phosphorus on copper concentration in the river.

stat.ME↗

arXiv2020

Deterministic Generation of Loss-Tolerant Photonic Cluster States with a Single Quantum Emitter

A photonic cluster state with a tree-type entanglement structure constitutes an efficient resource for quantum error correction of photon loss. But the generation of a tree cluster state with an arbitrary size is notoriously difficult. Here, we propose a protocol to deterministically generate photonic tree states of arbitrary size by using only a single quantum emitter. Photonic entanglement is established through both emission and re-scattering from the same emitter, enabling fast and resource-efficient entanglement generation. The same protocol can also be extended to generate more general tree-type entangled states.

quant-ph↗

arXiv2020

Narrow-linewidth tin-vacancy centers in a diamond waveguide

Integrating solid-state quantum emitters with photonic circuits is essential for realizing large-scale quantum photonic processors. Negatively charged tin-vacancy (SnV$^-$) centers in diamond have emerged as promising candidates for quantum emitters because of their excellent optical and spin properties including narrow-linewidth emission and long spin coherence times. SnV$^-$ centers need to be incorporated in optical waveguides for efficient on-chip routing of the photons they generate. However, such integration has yet to be realized. In this Letter, we demonstrate the coupling of SnV$^-$ centers to a nanophotonic waveguide. We realize this device by leveraging our recently developed shallow ion implantation and growth method for generation of high-quality SnV$^-$ centers and the advanced quasi-isotropic diamond fabrication technique. We confirm the compatibility and robustness of these techniques through successful coupling of narrow-linewidth SnV$^-$ centers (as narrow as $36\pm2$ MHz) to the diamond waveguide. Furthermore, we investigate the stability of waveguide-coupled SnV$^-$ centers under resonant excitation. Our results are an important step toward SnV$^-$-based on-chip spin-photon interfaces, single-photon nonlinearity, and photon-mediated spin interactions.

physics.optics↗

arXiv2020

Spectrally reconfigurable quantum emitters enabled by optimized fast modulation

The ability to shape photon emission facilitates strong photon-mediated interactions between disparate physical systems, thereby enabling applications in quantum information processing, simulation and communication. Spectral control in solid state platforms such as color centers, rare earth ions, and quantum dots is particularly attractive for realizing such applications on-chip. Here we propose the use of frequency-modulated optical transitions for spectral engineering of single photon emission. Using a scattering-matrix formalism, we find that a two-level system, when modulated faster than its optical lifetime, can be treated as a single-photon source with a widely reconfigurable photon spectrum that is amenable to standard numerical optimization techniques. To enable the experimental demonstration of this spectral control scheme, we investigate the Stark tuning properties of the silicon vacancy in silicon carbide, a color center with promise for optical quantum information processing technologies. We find that the silicon vacancy possesses excellent spectral stability and tuning characteristics, allowing us to probe its fast modulation regime, observe the theoretically-predicted two-photon correlations, and demonstrate spectral engineering. Our results suggest that frequency modulation is a powerful technique for the generation of new light states with unprecedented control over the spectral and temporal properties of single photons.

quant-ph↗