arXiv subjects

Yuxuan Chen

Publications and source records attributed to Yuxuan Chen.

155 records · Page 9Linked to original sources

arXiv2021

Optimize Neural Fictitious Self-Play in Regret Minimization Thinking

Optimization of deep learning algorithms to approach Nash Equilibrium remains a significant problem in imperfect information games, e.g. StarCraft and poker. Neural Fictitious Self-Play (NFSP) has provided an effective way to learn approximate Nash Equilibrium without prior domain knowledge in imperfect information games. However, optimality gap was left as an optimization problem of NFSP and by solving the problem, the performance of NFSP could be improved. In this study, focusing on the optimality gap of NFSP, we have proposed a new method replacing NFSP's best response computation with regret matching method. The new algorithm can make the optimality gap converge to zero as it iterates, thus converge faster than original NFSP. We have conduct experiments on three typical environments of perfect-information games and imperfect information games in OpenSpiel and all showed that our new algorithm performances better than original NFSP.

cs.AI↗

arXiv2021

PiP: Planning-informed Trajectory Prediction for Autonomous Driving

It is critical to predict the motion of surrounding vehicles for self-driving planning, especially in a socially compliant and flexible way. However, future prediction is challenging due to the interaction and uncertainty in driving behaviors. We propose planning-informed trajectory prediction (PiP) to tackle the prediction problem in the multi-agent setting. Our approach is differentiated from the traditional manner of prediction, which is only based on historical information and decoupled with planning. By informing the prediction process with the planning of ego vehicle, our method achieves the state-of-the-art performance of multi-agent forecasting on highway datasets. Moreover, our approach enables a novel pipeline which couples the prediction and planning, by conditioning PiP on multiple candidate trajectories of the ego vehicle, which is highly beneficial for autonomous driving in interactive scenarios.

cs.CV↗

arXiv2020

Query-Key Normalization for Transformers

Low-resource language translation is a challenging but socially valuable NLP task. Building on recent work adapting the Transformer's normalization to this setting, we propose QKNorm, a normalization technique that modifies the attention mechanism to make the softmax function less prone to arbitrary saturation without sacrificing expressivity. Specifically, we apply $\ell_2$ normalization along the head dimension of each query and key matrix prior to multiplying them and then scale up by a learnable parameter instead of dividing by the square root of the embedding dimension. We show improvements averaging 0.928 BLEU over state-of-the-art bilingual benchmarks for 5 low-resource translation pairs from the TED Talks corpus and IWSLT'15.

cs.CL↗

arXiv2019

SynFuzz: Efficient Concolic Execution via Branch Condition Synthesis

Concolic execution is a powerful program analysis technique for exploring execution paths in a systematic manner. Compare to random-mutation-based fuzzing, concolic execution is especially good at exploring paths that are guarded by complex and tight branch predicates (e.g., (a*b) == 0xdeadbeef). The drawback, however, is that concolic execution engines are much slower than native execution. One major source of the slowness is that concolic execution engines have to the interpret instructions to maintain the symbolic expression of program variables. In this work, we propose SynFuzz, a novel approach to perform scalable concolic execution. SynFuzz achieves this goal by replacing interpretation with dynamic taint analysis and program synthesis. In particular, to flip a conditional branch, SynFuzz first uses operation-aware taint analysis to record a partial expression (i.e., a sketch) of its branch predicate. Then it uses oracle-guided program synthesis to reconstruct the symbolic expression based on input-output pairs. The last step is the same as traditional concolic execution - SynFuzz consults a SMT solver to generate an input that can flip the target branch. By doing so, SynFuzz can achieve an execution speed that is close to fuzzing while retain concolic execution's capability of flipping complex branch predicates. We have implemented a prototype of SynFuzz and evaluated it with three sets of programs: real-world applications, the LAVA-M benchmark, and the Google Fuzzer Test Suite (FTS). The evaluation results showed that SynFuzz was much more scalable than traditional concolic execution engines, was able to find more bugs in LAVA-M than most state-of-the-art concolic execution engine (QSYM), and achieved better code coverage on real-world applications and FTS.

cs.CR↗

arXiv2018

A Chip-scale, Full-Stokes Polarimeter

The polarization of light conveys unique information that can be exploited by crucial applications. The bulky and costly discrete optical components used in conventional polarimeters limit their broad adoption. A compact, low-cost polarimeter would bring this functionality into a myriad of new scenarios and revolutionize its exploitation. Here we present a high-performance, full-Stokes polarimeter on a silicon chip. A surface polarization splitter and on-chip optical interferometer circuit produce the analysis matrix of an optimally conditioned polarimeter. This solid-state polarimeter is a system-on-a-chip with exceptional compactness, stability, and speed that could be used singly or in integrated arrays. Large arrays can increase the speed and resolution of full-Stokes imaging; therefore, our design provides a scalable polarimeter solution.

physics.app-ph↗

arXiv2018

CommanderSong: A Systematic Approach for Practical Adversarial Voice Recognition

The popularity of ASR (automatic speech recognition) systems, like Google Voice, Cortana, brings in security concerns, as demonstrated by recent attacks. The impacts of such threats, however, are less clear, since they are either less stealthy (producing noise-like voice commands) or requiring the physical presence of an attack device (using ultrasound). In this paper, we demonstrate that not only are more practical and surreptitious attacks feasible but they can even be automatically constructed. Specifically, we find that the voice commands can be stealthily embedded into songs, which, when played, can effectively control the target system through ASR without being noticed. For this purpose, we developed novel techniques that address a key technical challenge: integrating the commands into a song in a way that can be effectively recognized by ASR through the air, in the presence of background noise, while not being detected by a human listener. Our research shows that this can be done automatically against real world ASR applications. We also demonstrate that such CommanderSongs can be spread through Internet (e.g., YouTube) and radio, potentially affecting millions of ASR users. We further present a new mitigation technique that controls this threat.

cs.CR↗

arXiv2018

Phase Coexistence for the Hard-Core Model on ${\mathbb Z}^2$

The hard-core model has attracted much attention across several disciplines, representing lattice gases in statistical physics and independent sets in discrete mathematics and computer science. On finite graphs, we are given a parameter $λ$, and an independent set $I$ arises with probability proportional to $λ^{|I|}$. On infinite graphs a Gibbs measure is defined as a suitable limit with the correct conditional probabilities, and we are interested in determining when this limit is unique and when there is phase coexistence, i.e., existence of multiple Gibbs measures. It has long been conjectured that on ${\mathbb Z}^2$ this model has a critical value $λ_c \approx 3.796$ with the property that if $λ< λ_c$ then it exhibits uniqueness of phase, while if $λ> λ_c$ then there is phase coexistence. Much of the work to date on this problem has focused on the regime of uniqueness, with the state of the art being recent work of Sinclair, Srivastava, Štefankovič and Yin showing that there is a unique Gibbs measure for all $λ< 2.538$. Here we give the first non-trivial result in the other direction, showing that there are multiple Gibbs measures for all $λ> 5.3506$. There is some potential for lowering this bound, but with the methods we are using we cannot hope to replace $5.3506$ with anything below about $4.8771$. Our proof begins along the lines of the standard Peierls argument, but we add two innovations. First, following ideas of Kotecký and Randall, we construct an event that distinguishes two boundary conditions and always has long contours associated with it, obviating the need to accurately enumerate short contours. Second, we obtain improved bounds on the number of contours by relating them to a new class of self-avoiding walks on an oriented version of ${\mathbb Z}^2$.

math.PR↗

arXiv2016

Band gap renormalization and work function tuning in MoSe2/hBN/Ru(0001) heterostructures

Here we report the successful growth of MoSe2 on single layer hexagonal boron nitride (hBN) on Ru(0001) substrate by using molecular beam epitaxy. We investigated the electronic structures of MoSe2 using scanning tunneling microscopy and spectroscopy. Surprisingly, we found that the quasi-particle gap of the MoSe2 on hBN/Ru is about 0.25 eV smaller than those on graphene or graphite substrates. We attribute this result to the strong interaction between hBN/Ru which causes residual metallic screening from the substrate. The surface of MoSe2 exhibits Moiré pattern that replicates the Moiré pattern of hBN/Ru. In addition, the electronic structure and the work function of MoSe2 are modulated electrostatically with an amplitude of ~ 0.13 eV. Most interestingly, this electrostatic modulation is spatially in phase with the Moiré pattern of hBN on Ru(0001) whose surface also exhibits a work function modulation of the same amplitude.

cond-mat.mtrl-sci↗

arXiv2015

Visualizing Band Offsets and Edge States in Bilayer-Monolayer Transition Metal Dichalcogenides Lateral Heterojunction

Semiconductor heterostructures are fundamental building blocks for many important device applications. The emergence of two-dimensional semiconductors opens up a new realm for creating heterostructures. As the bandgaps of transition metal dichalcogenides thin films have sensitive layer dependence, it is natural to create lateral heterojunctions using the same materials with different thicknesses. Using scanning tunneling microscopy and spectroscopy, here we show the real space image of electronic structures across the bilayer-monolayer interface in MoSe2 and WSe2. Most bilayer-monolayer heterojunctions are found to have a zigzag-orientated interface, and the band alignment of such atomically sharp heterojunctions is of type-I with a well-defined interface mode which acts as a narrower-gap quantum wire. The ability to utilize such commonly existing thickness terrace as lateral heterojunctions is a crucial addition to the tool set for device applications based on atomically thin transition metal dichalcogenides, with the advantage of easy and flexible implementation.

cond-mat.mtrl-sci↗

arXiv2015

Examples of Complete Solvability of 2D Classical Superintegrable Systems

Classical (maximal) superintegrable systems in $n$ dimensions are Hamiltonian systems with $2n-1$ independent constants of the motion, globally defined, the maximum number possible. They are very special because they can be solved algebraically. In this paper we show explicitly, mostly through examples of 2nd order superintegrable systems in 2 dimensions, how the trajectories can be determined in detail using rather elementary algebraic, geometric and analytic methods applied to the closed quadratic algebra of symmetries of the system. We treat a family of 2nd order degenerate systems: oscillator analogies on Darboux, nonzero constant curvature, and flat spaces, related to one another via contractions, and obeying Kepler's laws. Then we treat two 2nd order nondegenerate systems, an analogy of a caged Coulomb problem on the 2-sphere and its contraction to a Euclidean space caged Coulomb problem. In all cases the symmetry algebra structure provides detailed information about the trajectories. An interesting example is the occurrence of ''metronome orbits'', trajectories confined to an arc rather than a loop, which are indicated clearly from the structure equations but might be overlooked using more traditional methods. We also treat the Post-Winternitz system, an example of a classical 4th order superintegrable system that cannot be solved using separation of variables. Finally we treat a superintegrable system, related to the addition theorem for elliptic functions, whose constants of the motion are only rational in the momenta, a system of special interest because its constants of the motion generate a closed polynomial algebra. This paper contains many new results but we have tried to present most of the materials in a fashion that is easily accessible to nonexperts, in order to provide entrée to superintegrablity theory.

math-ph↗

arXiv2015

Probing critical point energies of transition metal dichalcogenides: surprising indirect gap of single layer $SL-WSe_2$

Understanding quasiparticle band structures of transition metal dichalcogenides (TMDs) is critical for technological advances of these materials for atomic layer electronics and photonics. Although theoretical calculations to date have shown qualitatively similar features, there exist subtle differences which can lead to important consequences in the device characteristics. For example, most calculations have shown that all single layer (SL) TMDs have direct band gaps, while some have shown that $SL-WSe_2$ have an indirect gap. Moreover, there are large variations in the reported quasiparticle gaps, corresponding to large variations in exciton binding energies. By using a comprehensive form of scanning tunneling spectroscopy, we have revealed detailed quasiparticle electronic structures in TMDs, including the quasi-particle gaps, critical point energy locations and their origins in the Brillouin Zones (BZs). We show that $SL-WSe_2$ actually has an indirect quasi-particle gap with the conduction band minimum located at the Q point (instead of K), albeit the two states are nearly degenerate. Its implications on optical properties are discussed. We have further observed rich quasi-particle electronic structures of TMDs as a function of atomic structures and spin-orbital couplings.

cond-mat.mes-hall↗