arXiv subjects

Tianyi Wang

Publications and source records attributed to Tianyi Wang.

144 records · Page 8Linked to original sources

arXiv2022

Diverse Imitation Learning via Self-Organizing Generative Models

Imitation learning is the task of replicating expert policy from demonstrations, without access to a reward function. This task becomes particularly challenging when the expert exhibits a mixture of behaviors. Prior work has introduced latent variables to model variations of the expert policy. However, our experiments show that the existing works do not exhibit appropriate imitation of individual modes. To tackle this problem, we adopt an encoder-free generative model for behavior cloning (BC) to accurately distinguish and imitate different modes. Then, we integrate it with GAIL to make the learning robust towards compounding errors at unseen states. We show that our method significantly outperforms the state of the art across multiple experiments.

cs.LG↗

arXiv2022

Which side are you on? Insider-Outsider classification in conspiracy-theoretic social media

Social media is a breeding ground for threat narratives and related conspiracy theories. In these, an outside group threatens the integrity of an inside group, leading to the emergence of sharply defined group identities: Insiders -- agents with whom the authors identify and Outsiders -- agents who threaten the insiders. Inferring the members of these groups constitutes a challenging new NLP task: (i) Information is distributed over many poorly-constructed posts; (ii) Threats and threat agents are highly contextual, with the same post potentially having multiple agents assigned to membership in either group; (iii) An agent's identity is often implicit and transitive; and (iv) Phrases used to imply Outsider status often do not follow common negative sentiment patterns. To address these challenges, we define a novel Insider-Outsider classification task. Because we are not aware of any appropriate existing datasets or attendant models, we introduce a labeled dataset (CT5K) and design a model (NP2IO) to address this task. NP2IO leverages pretrained language modeling to classify Insiders and Outsiders. NP2IO is shown to be robust, generalizing to noun phrases not seen during training, and exceeding the performance of non-trivial baseline models by $20\%$.

cs.CL↗

arXiv2021

Realized GARCH, CBOE VIX, and the Volatility Risk Premium

We show that the Realized GARCH model yields close-form expression for both the Volatility Index (VIX) and the volatility risk premium (VRP). The Realized GARCH model is driven by two shocks, a return shock and a volatility shock, and these are natural state variables in the stochastic discount factor (SDF). The volatility shock endows the exponentially affine SDF with a compensation for volatility risk. This leads to dissimilar dynamic properties under the physical and risk-neutral measures that can explain time-variation in the VRP. In an empirical application with the S&P 500 returns, the VIX, and the VRP, we find that the Realized GARCH model significantly outperforms conventional GARCH models.

econ.EM↗

arXiv2021

The negative dependence of evacuation time on group size under a binding mechanism

This paper initiates the analysis of the relation between evacuation time and group size by applying an extended floor field cellular automaton model. Agents with various speeds, a group structure containing leaders and followers, and a dynamic field dependent on local population density are implemented all together in the model. Most importantly, a complete binding mechanism which includes leaders waiting for followers is brought up for the first time. A counterintuitive negative relation between evacuation time and group size is discovered in simulations. An entropy like quantity, namely the mixing index, is constructed to analyze the cause of that relation. It is found that under the binding mechanism, the higher degree of group mixing, the longer the evacuation time will be. Moreover, through a constant scale transformation, it is shown that the mixing index can be a key indicator that contains useful information about the evacuation system.

nlin.CG↗

arXiv2021

An Underwater Image Semantic Segmentation Method Focusing on Boundaries and a Real Underwater Scene Semantic Segmentation Dataset

With the development of underwater object grabbing technology, underwater object recognition and segmentation of high accuracy has become a challenge. The existing underwater object detection technology can only give the general position of an object, unable to give more detailed information such as the outline of the object, which seriously affects the grabbing efficiency. To address this problem, we label and establish the first underwater semantic segmentation dataset of real scene(DUT-USEG:DUT Underwater Segmentation Dataset). The DUT- USEG dataset includes 6617 images, 1487 of which have semantic segmentation and instance segmentation annotations, and the remaining 5130 images have object detection box annotations. Based on this dataset, we propose a semi-supervised underwater semantic segmentation network focusing on the boundaries(US-Net: Underwater Segmentation Network). By designing a pseudo label generator and a boundary detection subnetwork, this network realizes the fine learning of boundaries between underwater objects and background, and improves the segmentation effect of boundary areas. Experiments show that the proposed method improves by 6.7% in three categories of holothurian, echinus, starfish in DUT-USEG dataset, and achieves state-of-the-art results. The DUT- USEG dataset will be released at https://github.com/baxiyi/DUT-USEG.

cs.CV↗

arXiv2021

Legal Judgment Prediction with Multi-Stage CaseRepresentation Learning in the Real Court Setting

Legal judgment prediction(LJP) is an essential task for legal AI. While prior methods studied on this topic in a pseudo setting by employing the judge-summarized case narrative as the input to predict the judgment, neglecting critical case life-cycle information in real court setting could threaten the case logic representation quality and prediction correctness. In this paper, we introduce a novel challenging dataset from real courtrooms to predict the legal judgment in a reasonably encyclopedic manner by leveraging the genuine input of the case -- plaintiff's claims and court debate data, from which the case's facts are automatically recognized by comprehensively understanding the multi-role dialogues of the court debate, and then learnt to discriminate the claims so as to reach the final judgment through multi-task learning. An extensive set of experiments with a large civil trial data set shows that the proposed model can more accurately characterize the interactions among claims, fact and debate for legal judgment prediction, achieving significant improvements over strong state-of-the-art baselines. Moreover, the user study conducted with real judges and law school students shows the neural predictions can also be interpretable and easily observed, and thus enhancing the trial efficiency and judgment quality.

cs.CL↗

arXiv2021

Hierarchically Modeling Micro and Macro Behaviors via Multi-Task Learning for Conversion Rate Prediction

Conversion Rate (\emph{CVR}) prediction in modern industrial e-commerce platforms is becoming increasingly important, which directly contributes to the final revenue. In order to address the well-known sample selection bias (\emph{SSB}) and data sparsity (\emph{DS}) issues encountered during CVR modeling, the abundant labeled macro behaviors ($i.e.$, user's interactions with items) are used. Nonetheless, we observe that several purchase-related micro behaviors ($i.e.$, user's interactions with specific components on the item detail page) can supplement fine-grained cues for \emph{CVR} prediction. Motivated by this observation, we propose a novel \emph{CVR} prediction method by Hierarchically Modeling both Micro and Macro behaviors ($HM^3$). Specifically, we first construct a complete user sequential behavior graph to hierarchically represent micro behaviors and macro behaviors as one-hop and two-hop post-click nodes. Then, we embody $HM^3$ as a multi-head deep neural network, which predicts six probability variables corresponding to explicit sub-paths in the graph. They are further combined into the prediction targets of four auxiliary tasks as well as the final $CVR$ according to the conditional probability rule defined on the graph. By employing multi-task learning and leveraging the abundant supervisory labels from micro and macro behaviors, $HM^3$ can be trained end-to-end and address the \emph{SSB} and \emph{DS} issues. Extensive experiments on both offline and online settings demonstrate the superiority of the proposed $HM^3$ over representative state-of-the-art methods.

cs.LG↗

arXiv2021

Liquidation, Leverage and Optimal Margin in Bitcoin Futures Markets

Using the generalized extreme value theory to characterize tail distributions, we address liquidation, leverage, and optimal margins for bitcoin long and short futures positions. The empirical analysis of perpetual bitcoin futures on BitMEX shows that (1) daily forced liquidations to out- standing futures are substantial at 3.51%, and 1.89% for long and short; (2) investors got forced liquidation do trade aggressively with average leverage of 60X; and (3) exchanges should elevate current 1% margin requirement to 33% (3X leverage) for long and 20% (5X leverage) for short to reduce the daily margin call probability to 1%. Our results further suggest normality assumption on return significantly underestimates optimal margins. Policy implications are also discussed.

q-fin.TR↗

arXiv2020

Accuracy and Resiliency of Analog Compute-in-Memory Inference Engines

Recently, analog compute-in-memory (CIM) architectures based on emerging analog non-volatile memory (NVM) technologies have been explored for deep neural networks (DNN) to improve energy efficiency. Such architectures, however, leverage charge conservation, an operation with infinite resolution, and thus are susceptible to errors. The computations in DNN realized by analog NVM thus have high uncertainty due to the device stochasticity. Several reports have demonstrated the use of analog NVM for CIM in a limited scale. It is unclear whether the uncertainties in computations will prohibit large-scale DNNs. To explore this critical issue of scalability, this paper first presents a simulation framework to evaluate the feasibility of large-scale DNNs based on CIM architecture and analog NVM. Simulation results show that DNNs trained for high-precision digital computing engines are not resilient against the uncertainty of the analog NVM devices. To avoid such catastrophic failures, this paper introduces the analog floating-point representation for the DNN, and the Hessian-Aware Stochastic Gradient Descent (HA-SGD) training algorithm to enhance the inference accuracy of trained DNNs. As a result of such enhancements, DNNs such as Wide ResNets for the CIFAR-100 image recognition problem are demonstrated to have significant performance improvements in accuracy without adding cost to the inference hardware.

eess.SP↗

arXiv2020

Towards full surface Brillouin zone mapping by coherent multi-photon photoemission

We report a novel approach for coherent multi-photon photoemission band mapping of the entire Brillouin zone with infrared light that is readily implemented in a laboratory setting. We excite a solid state material, Ag(110), with intense femtosecond laser pulses to excite higher-order multi-photon photoemission; angle-resolved electron spectroscopic acquisition records photoemission at large in-plane momenta involving optical transitions from the occupied to unoccupied bands of the sample that otherwise might remain hidden by the photoemission horizon. We propose this as a complementary ultrafast method to time- and angle-resolved two-color, e.g. infrared pump and extreme ultraviolet probe, photoemission spectroscopy, with the advantage of being able to measure and control the coherent electron dynamics.

cond-mat.other↗

arXiv2020

Masking Orchestration: Multi-task Pretraining for Multi-role Dialogue Representation Learning

Multi-role dialogue understanding comprises a wide range of diverse tasks such as question answering, act classification, dialogue summarization etc. While dialogue corpora are abundantly available, labeled data, for specific learning tasks, can be highly scarce and expensive. In this work, we investigate dialogue context representation learning with various types unsupervised pretraining tasks where the training objectives are given naturally according to the nature of the utterance and the structure of the multi-role conversation. Meanwhile, in order to locate essential information for dialogue summarization/extraction, the pretraining process enables external knowledge integration. The proposed fine-tuned pretraining mechanism is comprehensively evaluated via three different dialogue datasets along with a number of downstream dialogue-mining tasks. Result shows that the proposed pretraining mechanism significantly contributes to all the downstream tasks without discrimination to different encoders.

cs.CL↗

arXiv2019

Quantification of Ion Migration in CH3NH3PbI3 Perovskite Solar Cells by Transient Capacitance Measurements

Solar cells based on organic-inorganic metal halide perovskites show efficiencies close to highly-optimized silicon solar cells. However, ion migration in the perovskite films leads to device degradation and impedes large scale commercial applications. We use transient ion-drift measurements to quantify activation energy, diffusion coefficient, and concentration of mobile ions in methylammonium lead triiodide (MAPbI3) perovskite solar cells, and find that their properties change close to the tetragonal-to-orthorhombic phase transition temperature. We identify three migrating ion species which we attribute to the migration of iodide (I-) and methylammonium (MA+). We find that the concentration of mobile MA+ ions is one order of magnitude higher than the one of mobile I- ions, and that the diffusion coefficient of mobile MA+ ions is three orders of magnitude lower than the one for mobile I- ions. We furthermore observe that the activation energy of mobile I- ions (0.29 eV) is highly reproducible for different devices, while the activation energy of mobile MA+ depends strongly on device fabrication. This quantification of mobile ions in MAPbI3 will lead to a better understanding of ion migration and its role in operation and degradation of perovskite solar cells.

cond-mat.mtrl-sci↗

arXiv2016

Indirect to direct bandgap transition in methylammonium lead halide perovskite

Methylammonium lead iodide perovskites are considered direct bandgap semiconductors. Here we show that in fact they present a weakly indirect bandgap 60 meV below the direct bandgap transition. This is a consequence of spin-orbit coupling resulting in Rashba-splitting of the conduction band. The indirect nature of the bandgap explains the apparent contradiction of strong absorption and long charge carrier lifetime. Under hydrostatic pressure from ambient to 325 MPa, Rashba splitting is reduced due to a pressure induced ordering of the crystal structure. The nature of the bandgap becomes increasingly more direct, resulting in five times faster charge carrier recombination, and a doubling of the radiative efficiency. At hydrostatic pressures above 325 MPa, MAPI undergoes a reversible phase transition resulting in a purely direct bandgap semiconductor. The pressure-induced changes suggest epitaxial and synthetic routes to higher efficiency optoelectronic devices.

cond-mat.mtrl-sci↗

arXiv2016

Store Location Selection via Mining Search Query Logs of Baidu Maps

Choosing a good location when opening a new store is crucial for the future success of a business. Traditional methods include offline manual survey, which is very time consuming, and analytic models based on census data, which are un- able to adapt to the dynamic market. The rapid increase of the availability of big data from various types of mobile devices, such as online query data and offline positioning data, provides us with the possibility to develop automatic and accurate data-driven prediction models for business store placement. In this paper, we propose a Demand Distribution Driven Store Placement (D3SP) framework for business store placement by mining search query data from Baidu Maps. D3SP first detects the spatial-temporal distributions of customer demands on different business services via query data from Baidu Maps, the largest online map search engine in China, and detects the gaps between demand and sup- ply. Then we determine candidate locations via clustering such gaps. In the final stage, we solve the location optimization problem by predicting and ranking the number of customers. We not only deploy supervised regression models to predict the number of customers, but also learn to rank models to directly rank the locations. We evaluate our framework on various types of businesses in real-world cases, and the experiments results demonstrate the effectiveness of our methods. D3SP as the core function for store placement has already been implemented as a core component of our business analytics platform and could be potentially used by chain store merchants on Baidu Nuomi.

cs.AI↗

arXiv2016

Defending against Sybil Devices in Crowdsourced Mapping Services

Real-time crowdsourced maps such as Waze provide timely updates on traffic, congestion, accidents and points of interest. In this paper, we demonstrate how lack of strong location authentication allows creation of software-based {\em Sybil devices} that expose crowdsourced map systems to a variety of security and privacy attacks. Our experiments show that a single Sybil device with limited resources can cause havoc on Waze, reporting false congestion and accidents and automatically rerouting user traffic. More importantly, we describe techniques to generate Sybil devices at scale, creating armies of virtual vehicles capable of remotely tracking precise movements for large user populations while avoiding detection. We propose a new approach to defend against Sybil devices based on {\em co-location edges}, authenticated records that attest to the one-time physical co-location of a pair of devices. Over time, co-location edges combine to form large {\em proximity graphs} that attest to physical interactions between devices, allowing scalable detection of virtual vehicles. We demonstrate the efficacy of this approach using large-scale simulations, and discuss how they can be used to dramatically reduce the impact of attacks against crowdsourced mapping services.

cs.SI↗

arXiv2014

Crowds on Wall Street: Extracting Value from Social Investing Platforms

For decades, the world of financial advisors has been dominated by large investment banks such as Goldman Sachs. In recent years, user-contributed investment services such as SeekingAlpha and StockTwits have grown to millions of users. In this paper, we seek to understand the quality and impact of content on social investment platforms, by empirically analyzing complete datasets of SeekingAlpha articles (9 years) and StockTwits messages (4 years). We develop sentiment analysis tools and correlate contributed content to the historical performance of relevant stocks. While SeekingAlpha articles and StockTwits messages provide minimal correlation to stock performance in aggregate, a subset of authors contribute more valuable (predictive) content. We show that these authors can be identified via both empirical methods or by user interactions, and investments using their analysis significantly outperform broader markets. Finally, we conduct a user survey that sheds light on users views of SeekingAlpha content and stock manipulation.

cs.SI↗

arXiv2011

On Multi-Dimensional Sonic-Subsonic Flow

In this paper, a compensated compactness framework is established for sonic-subsonic approximate solutions to the $n$-dimensional$(n\geq 2)$ Euler equations for steady irrotational flow that may contain stagnation points. This compactness framework holds provided that the approximate solutions are uniformly bounded and satisfy $H^{-1}_{loc}(Ω)$ compactness conditions. As illustration, we show the existence of sonic-subsonic weak solution to n-dimensional$(n\geq 2)$ Euler equations for steady irrotational flow past obstacles or through an infinitely long nozzle. This is the first result concerning the sonic-subsonic limit for $n$-dimension$(n\geq 3)$.

math.AP↗

arXiv2010

Vanishing Viscosity Limit for Isentropic Navier-Stokes Equations with Density-dependent Viscosity

In this paper, we study the vanishing viscosity limit of one-dimensional isentropic compressible Navier-Stokes equations with density-dependent viscosity, to the isentropic compressible Euler equations. Based on several new uniform estimates to the viscous systems, in addition to the framework recently established by G. Chen and M. Perepelitsa, we justify that the finite energy solution of the isentropic compressible Euler equations for a large class of initial data can be obtained as the inviscid limit of the compressible Navier-Stokes equations even when the viscosity depends on the density.

math.AP↗