Searcharxiv⌕ Search

arXiv subjects

Yi Huang

Publications and source records attributed to Yi Huang.

At least 181 records · Page 10Linked to original sources
arXiv2022

"Think Before You Speak": Improving Multi-Action Dialog Policy by Planning Single-Action Dialogs

Multi-action dialog policy (MADP), which generates multiple atomic dialog actions per turn, has been widely applied in task-oriented dialog systems to provide expressive and efficient system responses. Existing MADP models usually imitate action combinations from the labeled multi-action dialog samples. Due to data limitations, they generalize poorly toward unseen dialog flows. While interactive learning and reinforcement learning algorithms can be applied to incorporate external data sources of real users and user simulators, they take significant manual effort to build and suffer from instability. To address these issues, we propose Planning Enhanced Dialog Policy (PEDP), a novel multi-task learning framework that learns single-action dialog dynamics to enhance multi-action prediction. Our PEDP method employs model-based planning for conceiving what to express before deciding the current response through simulating single-action dialogs. Experimental results on the MultiWOZ dataset demonstrate that our fully supervised learning-based method achieves a solid task success rate of 90.6%, improving 3% compared to the state-of-the-art methods.

cs.CL↗
arXiv2022

Coordinated Power Control for Network Integrated Sensing and Communication

This correspondence paper studies a network integrated sensing and communication (ISAC) system that unifies the interference channel for communication and distributed radar sensing. In this system, a set of distributed ISAC transmitters send individual messages to their respective communication users (CUs), and at the same time cooperate with multiple sensing receivers to estimate the location of one target. We exploit the coordinated power control among ISAC transmitters to minimize their total transmit power while ensuring the minimum signal-to-interference-plus-noise ratio (SINR) constraints at individual CUs and the maximum Cramér-Rao lower bound (CRLB) requirement for target location estimation. Although the formulated coordinated power control problem is non-convex and difficult to solve in general, we propose two efficient algorithms to obtain high-quality solutions based on the semi-definite relaxation (SDR) and CRLB approximation, respectively. Numerical results show that the proposed designs achieve substantial performance gains in terms of power reduction, as compared to the benchmark with a heuristic separate communication-sensing design.

cs.IT↗
arXiv2022

Solving Simulation Systematics in and with AI/ML

Training an AI/ML system on simulated data while using that system to infer on data from real detectors introduces a systematic error which is difficult to estimate and in many analyses is simply not confronted. It is crucial to minimize and to quantitatively estimate the uncertainties in such analysis and do so with a precision and accuracy that matches those that AI/ML techniques bring. Here we highlight the need to confront this class of systematic error, discuss conventional ways to estimate it and describe ways to quantify and to minimize the uncertainty using methods which are themselves based on the power of AI/ML. We also describe methods to introduce a simulation into an AI/ML network to allow for training of its semantically meaningful parameters. This whitepaper is a contribution to the Computational Frontier of Snowmass21.

hep-ex↗
arXiv2022

Conductivity of two-dimensional narrow gap semiconductors subjected to strong Coulomb disorder

In the ideal disorder-free situation, a two-dimensional band gap insulator has an activation energy for conductivity equal to half the band gap $Δ$. But transport experiments usually exhibit a much smaller activation energy at low temperature, and the relation between this activation energy and $Δ$ is unclear. Here we consider the temperature-dependent conductivity of a two-dimensional insulator on a substrate containing Coulomb impurities, with random potential amplitude $Γ\gg Δ$. We show that the conductivity generically exhibits three regimes of conductivity, and only the highest temperature regime exhibits an activation energy that reflects the band gap. At lower temperatures, the conduction proceeds through activated hopping or Efros-Shklovskii variable-range hopping between electron and hole puddles created by the disorder. We show that the activation energy and characteristic temperature associated with these processes steeply collapse near a critical impurity concentration. Larger concentrations lead to an exponentially small activation energy and exponentially long localization length, which in mesoscopic samples can appear as a disorder-induced insulator-to-metal transition. We also arrive at a similar steep disorder driven insulator-metal transition in thin films of three-dimensional topological insulators with large dielectric constant, for which Coulomb impurities inside the film create a large disorder potential due to confinement of their electric field inside the film.

cond-mat.mes-hall↗
arXiv2022

Passive Intermodulation at Contacts of Rough Conductors

Passive intermodulation (PIM) is a niggling phenomenon that debilitates performance of modern communications and navigation systems. PIM products interfere with the information signals and cause their nonlinear distortion. The sources and basic mechanisms of PIM were studied in literature but PIM remains a serious problem of signal integrity. In this paper, the main sources and mechanisms of PIM generation by joints of good conductors are discussed. It is shown that the passive electrical, thermal and mechanical nonlinearities are intrinsically linked despite their distinctively different time scales. The roughness of the contact surfaces plays an important role in PIM generation by conductor joints. A review of the PIM phenomenology at contacts of the good conductors suggests that novel multiphysics models are necessary for the analysis and reliable prediction of PIM products generated by several concurrent nonlinearities of diverse physical nature.

physics.app-ph↗
arXiv2022

High fill factor confocal compound eyes fabricated by direct laser writing for better imaging quality

We fabricate two kinds of 100% fill factor compound eye structures using direct laser writing, including conventional compound eyes (CVCEs) with the same focal length of each microlens unit, and specially designed confocal compound eyes (CFCEs). For CFCEs, the focal length of each microlens unit is determined by its position and is equal to the distance between the microlens unit and the image sensor. In this letter, the optical properties of CVCEs and CFCEs are tested and compared. It is found that compared with CVCEs, CFCEs can improve the focusing efficiency by about 7%, enlarge the imaging area by about 25%, and have better imaging quality at the edge of the field of view.

physics.optics↗
arXiv2021

High Energy Irradiation Effects on Silicon Photonic Passive Devices

In this work, the radiation responses of silicon photonic passive devices built in silicon-on-insulator (SOI) technology are investigated through high energy neutron and 60Co gamma-ray irradiation. The wavelengths of both micro-ring resonators (MRRs) and Mach-Zehnder interferometers (MZIs) exhibit blue shifts after high-energy neutron irradiation to a fluence of 1*1012 n/cm2; the blue shift is smaller in MZI devices than in MRRs due to different waveguide widths. Devices with SiO2 upper cladding layer show strong tolerance to irradiation. Neutron irradiation leads to slight changes in the crystal symmetry in the Si cores of the optical devices and accelerated oxidization for devices without SiO2 cladding. A 2 um top cladding of SiO2 layer significantly improves the radiation tolerance of these passive photonic devices.

physics.optics↗
arXiv2021

CAMA: Energy and Memory Efficient Automata Processing in Content-Addressable Memories

Accelerating finite automata processing is critical for advancing real-time analytic in pattern matching, data mining, bioinformatics, intrusion detection, and machine learning. Recent in-memory automata accelerators leveraging SRAMs and DRAMs have shown exciting improvements over conventional digital designs. However, the bit-vector representation of state transitions used by all SOTA designs is only optimal in processing worst-case completely random patterns, while a significant amount of memory and energy is wasted in running most real-world benchmarks. We present CAMA, a Content-Addressable Memory (CAM) enabled Automata accelerator for processing homogeneous non-deterministic finite automata (NFA). A radically different state representation scheme, along with co-designed novel circuits and data encoding schemes, greatly reduces energy, memory, and chip area for most realistic NFAs. CAMA is holistically optimized with the following major contributions: (1) a 16x256 8-transistor (8T) CAM array for state matching, replacing the 256x256 6T SRAM array or two 16x256 6T SRAM banks in SOTA designs; (2) a novel encoding scheme that enables content searching within 8T SRAMs and adapts to different applications; (3) a reconfigurable and scalable architecture that improves efficiency on all tested benchmarks, without losing support for any NFA that is compatible with SOTA designs; (4) an optimization framework that automates the choice of encoding schemes and maps a given NFA to the proposed hardware. Two versions of CAMA, one optimized for energy (CAMA-E) and the other for throughput (CAMA-T), are comprehensively evaluated in a 28nm CMOS process, and across 21 real-world and synthetic benchmarks. CAMA-E achieves 2.1x, 2.8x, and 2.04x lower energy than CA, 2-stride Impala, and eAP. CAMA-T shows 2.68x, 3.87x and 2.62x higher average compute density than 2-stride Impala, CA, and eAP.

cs.AR↗
arXiv2021

Extraction of Medication Names from Twitter Using Augmentation and an Ensemble of Language Models

The BioCreative VII Track 3 challenge focused on the identification of medication names in Twitter user timelines. For our submission to this challenge, we expanded the available training data by using several data augmentation techniques. The augmented data was then used to fine-tune an ensemble of language models that had been pre-trained on general-domain Twitter content. The proposed approach outperformed the prior state-of-the-art algorithm Kusuri and ranked high in the competition for our selected objective function, overlapping F1 score.

cs.CL↗
arXiv2021

Efficient Data Compression for 3D Sparse TPC via Bicephalous Convolutional Autoencoder

Real-time data collection and analysis in large experimental facilities present a great challenge across multiple domains, including high energy physics, nuclear physics, and cosmology. To address this, machine learning (ML)-based methods for real-time data compression have drawn significant attention. However, unlike natural image data, such as CIFAR and ImageNet that are relatively small-sized and continuous, scientific data often come in as three-dimensional data volumes at high rates with high sparsity (many zeros) and non-Gaussian value distribution. This makes direct application of popular ML compression methods, as well as conventional data compression methods, suboptimal. To address these obstacles, this work introduces a dual-head autoencoder to resolve sparsity and regression simultaneously, called \textit{Bicephalous Convolutional AutoEncoder} (BCAE). This method shows advantages both in compression fidelity and ratio compared to traditional data compression methods, such as MGARD, SZ, and ZFP. To achieve similar fidelity, the best performer among the traditional methods can reach only half the compression ratio of BCAE. Moreover, a thorough ablation study of the BCAE method shows that a dedicated segmentation decoder improves the reconstruction.

cs.LG↗
arXiv2021

Balancing Methods for Multi-label Text Classification with Long-Tailed Class Distribution

Multi-label text classification is a challenging task because it requires capturing label dependencies. It becomes even more challenging when class distribution is long-tailed. Resampling and re-weighting are common approaches used for addressing the class imbalance problem, however, they are not effective when there is label dependency besides class imbalance because they result in oversampling of common labels. Here, we introduce the application of balancing loss functions for multi-label text classification. We perform experiments on a general domain dataset with 90 labels (Reuters-21578) and a domain-specific dataset from PubMed with 18211 labels. We find that a distribution-balanced loss function, which inherently addresses both the class imbalance and label linkage problems, outperforms commonly used loss functions. Distribution balancing methods have been successfully used in the image recognition field. Here, we show their effectiveness in natural language processing. Source code is available at https://github.com/Roche/BalancedLossNLP.

cs.CL↗
arXiv2021

Super McShane identity

The authors derive a McShane identity for once-punctured super tori. Relying upon earlier work on super Teichmüller theory by the last two-named authors, they further develop the supergeometry of these surfaces and establish asymptotic growth rate of their length spectra.

math.GT↗
arXiv2021

Disorder effects in topological insulator nanowires

Three-dimensional topological insulator (TI) nanowires with quantized surface subband spectra are studied as a main component of Majorana bound states (MBS) devices. However, such wires are known to have large concentration $N \sim 10^{19}$ cm$^{-3}$ of Coulomb impurities. It is believed that a MBS device can function only if the amplitude of long-range fluctuations of the random Coulomb potential $Γ$ is smaller than the subband gap $Δ$. Here we calculate $Γ$ for recently experimentally studied large-dielectric-constant (Bi$_{1-x}$Sb$_x$)$_2$Te$_{3}$ wires in a small-dielectric-constant environment (no superconductor). We show that provided by such a dielectric-constant contrast, the confinement of electric field of impurities within the wire allows more distant impurities to contribute into $Γ$, leading to $Γ\sim 3Δ$. We also calculate a TI wire resistance as a function of the Fermi level and carrier concentration due to scattering on Coulomb and neutral impurities, and do not find observable discrete subband-spectrum related oscillations at $N \gtrsim 10^{18}$ cm$^{-3}$.

cond-mat.mes-hall↗
arXiv2021

A single-frequency single-resonator laser on erbium-doped lithium niobate on insulator

Erbium-doped lithium niobate on insulator (Er:LNOI) is a promising platform for photonic integrated circuits as it adds gain to the LNOI system and enables on-chip lasers and amplifiers. A challenge for Er:LNOI laser is to increase its output power while maintaining single-frequency and single (-transverse)-mode operation. In this work, we demonstrate that single-frequency and single-mode operation can be achieved even in a single multi-mode Er:LNOI microring by introducing mode-dependent loss and gain competition. In a single microring with a free spectral range of 192 GHz, we have achieved single-mode lasing with an output power of 2.1 microwatt, a side-mode suppression of 35.5 dB, and a linewidth of 1.27 MHz.

physics.optics↗
arXiv2021

Metal-Insulator Transition in $n$-type bulk crystals and films of strongly compensated SrTiO$_3$

We start by analyzing experimental data of Spinelli [A. Spinelli, M. A. Torija, C. Liu, C. Jan, and C. Leighton, Phys. Rev. B 81, 155110 (2010)] for conductivity of $n$-type bulk crystals of SrTiO$_3$ (STO) with broad electron concentration $n$ range of $4\times 10^{15}$ - $4 \times10^{20} $ cm$^{-3}$, at low temperatures. We obtain good fit of the conductivity data, $σ(n)$, by the Drude formula for $n \geq n_c \simeq 3 \times 10^{16} $ cm$^{-3}$ assuming that used for doping insulating STO bulk crystals are strongly compensated and the total concentration of background charged impurities is $N = 10^{19}$ cm$^{-3}$. At $n< n_c$, the conductivity collapses with decreasing $n$ and the Drude theory fit fails. We argue that this is the metal-insulator transition (MIT) in spite of the very large Bohr radius of hydrogen-like donor state $a_B \simeq 700$ nm with which the Mott criterion of MIT for a weakly compensated semiconductor, $na_B^3 \simeq 0.02$, predicts $10^{5}$ times smaller $n_c$. We try to explain this discrepancy in the framework of the theory of the percolation MIT in a strongly compensated semiconductor with the same $N=10^{19}$ cm$^{-3}$. In the second part of this paper, we develop the percolation MIT theory for films of strongly compensated semiconductors. We apply this theory to doped STO films with thickness $d \leq 130$ nm and calculate the critical MIT concentration $n_c(d)$. We find that, for doped STO films on insulating STO bulk crystals, $n_c(d)$ grows with decreasing $d$. Remarkably, STO films in a low dielectric constant environment have the same $n_c(d)$. This happens due to the Rytova-Keldysh modification of a charge impurity potential which allows a larger number of the film charged impurities to contribute to the random potential.

cond-mat.mtrl-sci↗
arXiv2021

Disorder effects in topological insulator thin films

Thin films of topological insulators (TI) attract large attention because of expected topological effects from the inter-surface hybridization of Dirac points. However, these effects may be depleted by unexpectedly large energy smearing $Γ$ of surface Dirac points by the random potential of abundant Coulomb impurities. We show that in a typical TI film with large dielectric constant $\sim 50$ sandwiched between two low dielectric constant layers, the Rytova-Chaplik-Entin-Keldysh modification of the Coulomb potential of a charge impurity allows a larger number of the film impurities to contribute to $Γ$. As a result, $Γ$ is large and independent of the TI film thickness $d$ for $d > 5$ nm. In thinner films $Γ$ grows with decreasing $d$ due to reduction of screening by the hybridization gap. We study the surface conductivity away from the neutrality point and at the neutrality point. In the latter case, we find the maximum TI film thickness at which the hybridization gap is still able to make a TI film insulating and allow observation of the quantum spin Hall effect, $d_{\max} \sim 7$ nm.

cond-mat.mes-hall↗
arXiv2021

Distributed algorithms for the least square solution of linear equations

This paper proposes distributed algorithms for solving linear equations to seek a least square solution via multi-agent networks. We consider that each agent has only access to a small and imcomplete block of linear equations rather than the complete row or column in the existing literatures. Firstly, we focus on the case of a homogeneous partition of linear equations. A distributed algorithm is proposed via a single-layered grid network, in which each agent only needs to control three scalar states. Secondly, we consider the case of heterogeneous partitions of linear equations. Two distributed algorithms with doubled-layered network are developed, which allows each agent's states to have different dimensions and can be applied to heterogeneous agents with different storage and computation capability. Rigorous proofs show that the proposed distributed algorithms collaboratively obtain a least square solution with exponential convergence, and also own a solvability verification property, i.e., a criterion to verify whether the obtained solution is an exact solution. Finally, some simulation examples are provided to demonstrate the effectiveness of the proposed algorithms.

math.NA↗