arXiv subjects

Cheng Zhang

Publications and source records attributed to Cheng Zhang.

At least 379 records · Page 21Linked to original sources

arXiv2021

Meta Sequence Learning for Generating Adequate Question-Answer Pairs

Creating multiple-choice questions to assess reading comprehension of a given article involves generating question-answer pairs (QAPs) on the main points of the document. We present a learning scheme to generate adequate QAPs via meta-sequence representations of sentences. A meta sequence is a sequence of vectors comprising semantic and syntactic tags. In particular, we devise a scheme called MetaQA to learn meta sequences from training data to form pairs of a meta sequence for a declarative sentence (MD) and a corresponding interrogative sentence (MIs). On a given declarative sentence, a trained MetaQA model converts it to a meta sequence, finds a matched MD, and uses the corresponding MIs and the input sentence to generate QAPs. We implement MetaQA for the English language using semantic-role labeling, part-of-speech tagging, and named-entity recognition, and show that trained on a small dataset, MetaQA generates efficiently over the official SAT practice reading tests a large number of syntactically and semantically correct QAPs with over 97\% accuracy.

cs.CL↗

arXiv2021

MosaicOS: A Simple and Effective Use of Object-Centric Images for Long-Tailed Object Detection

Many objects do not appear frequently enough in complex scenes (e.g., certain handbags in living rooms) for training an accurate object detector, but are often found frequently by themselves (e.g., in product images). Yet, these object-centric images are not effectively leveraged for improving object detection in scene-centric images. In this paper, we propose Mosaic of Object-centric images as Scene-centric images (MosaicOS), a simple and novel framework that is surprisingly effective at tackling the challenges of long-tailed object detection. Keys to our approach are three-fold: (i) pseudo scene-centric image construction from object-centric images for mitigating domain differences, (ii) high-quality bounding box imputation using the object-centric images' class labels, and (iii) a multi-stage training procedure. On LVIS object detection (and instance segmentation), MosaicOS leads to a massive 60% (and 23%) relative improvement in average precision for rare object categories. We also show that our framework can be compatibly used with other existing approaches to achieve even further gains. Our pre-trained models are publicly available at https://github.com/czhang0528/MosaicOS/.

cs.CV↗

arXiv2021

DoWhy: Addressing Challenges in Expressing and Validating Causal Assumptions

Estimation of causal effects involves crucial assumptions about the data-generating process, such as directionality of effect, presence of instrumental variables or mediators, and whether all relevant confounders are observed. Violation of any of these assumptions leads to significant error in the effect estimate. However, unlike cross-validation for predictive models, there is no global validator method for a causal estimate. As a result, expressing different causal assumptions formally and validating them (to the extent possible) becomes critical for any analysis. We present DoWhy, a framework that allows explicit declaration of assumptions through a causal graph and provides multiple validation tests to check a subset of these assumptions. Our experience with DoWhy highlights a number of open questions for future research: developing new ways beyond causal graphs to express assumptions, the role of causal discovery in learning relevant parts of the graph, and developing validation tests that can better detect errors, both for average and conditional treatment effects. DoWhy is available at https://github.com/microsoft/dowhy.

cs.LG↗

arXiv2021

DeepPanoContext: Panoramic 3D Scene Understanding with Holistic Scene Context Graph and Relation-based Optimization

Panorama images have a much larger field-of-view thus naturally encode enriched scene context information compared to standard perspective images, which however is not well exploited in the previous scene understanding methods. In this paper, we propose a novel method for panoramic 3D scene understanding which recovers the 3D room layout and the shape, pose, position, and semantic category for each object from a single full-view panorama image. In order to fully utilize the rich context information, we design a novel graph neural network based context model to predict the relationship among objects and room layout, and a differentiable relationship-based optimization module to optimize object arrangement with well-designed objective functions on-the-fly. Realizing the existing data are either with incomplete ground truth or overly-simplified scene, we present a new synthetic dataset with good diversity in room layout and furniture placement, and realistic image quality for total panoramic 3D scene understanding. Experiments demonstrate that our method outperforms existing methods on panoramic scene understanding in terms of both geometry accuracy and object arrangement. Code is available at https://chengzhag.github.io/publication/dpc.

cs.CV↗

arXiv2021

Holistic 3D Scene Understanding from a Single Image with Implicit Representation

We present a new pipeline for holistic 3D scene understanding from a single image, which could predict object shapes, object poses, and scene layout. As it is a highly ill-posed problem, existing methods usually suffer from inaccurate estimation of both shapes and layout especially for the cluttered scene due to the heavy occlusion between objects. We propose to utilize the latest deep implicit representation to solve this challenge. We not only propose an image-based local structured implicit network to improve the object shape estimation, but also refine the 3D object pose and scene layout via a novel implicit scene graph neural network that exploits the implicit local object features. A novel physical violation loss is also proposed to avoid incorrect context between objects. Extensive experiments demonstrate that our method outperforms the state-of-the-art methods in terms of object shape, scene layout estimation, and 3D object detection.

cs.CV↗

arXiv2021

Toward Integrated Human-machine Intelligence for Civil Engineering: An Interdisciplinary Perspective

The purpose of this paper is to examine the opportunities and barriers of Integrated Human-Machine Intelligence (IHMI) in civil engineering. Integrating artificial intelligence's high efficiency and repeatability with humans' adaptability in various contexts can advance timely and reliable decision-making during civil engineering projects and emergencies. Successful cases in other domains, such as biomedical science, healthcare, and transportation, showed the potential of IHMI in data-driven, knowledge-based decision-making in numerous civil engineering applications. However, whether the industry and academia are ready to embrace the era of IHMI and maximize its benefit to the industry is still questionable due to several knowledge gaps. This paper thus calls for future studies in exploring the value, method, and challenges of applying IHMI in civil engineering. Our systematic review of the literature and motivating cases has identified four knowledge gaps in achieving effective IHMI in civil engineering. First, it is unknown what types of tasks in the civil engineering domain can be assisted by AI and to what extent. Second, the interface between human and AI in civil engineering-related tasks need more precise and formal definition. Third, the barriers that impede collecting detailed behavioral data from humans and contextual environments deserve systematic classification and prototyping. Lastly, it is unknown what expected and unexpected impacts will IHMI have on the AEC industry and entrepreneurship. Analyzing these knowledge gaps led to a list of identified research questions. This paper will lay the foundation for identifying relevant studies to form a research roadmap to address the four knowledge gaps identified.

cs.AI↗

arXiv2021

Countering Adversarial Examples: Combining Input Transformation and Noisy Training

Recent studies have shown that neural network (NN) based image classifiers are highly vulnerable to adversarial examples, which poses a threat to security-sensitive image recognition task. Prior work has shown that JPEG compression can combat the drop in classification accuracy on adversarial examples to some extent. But, as the compression ratio increases, traditional JPEG compression is insufficient to defend those attacks but can cause an abrupt accuracy decline to the benign images. In this paper, with the aim of fully filtering the adversarial perturbations, we firstly make modifications to traditional JPEG compression algorithm which becomes more favorable for NN. Specifically, based on an analysis of the frequency coefficient, we design a NN-favored quantization table for compression. Considering compression as a data augmentation strategy, we then combine our model-agnostic preprocess with noisy training. We fine-tune the pre-trained model by training with images encoded at different compression levels, thus generating multiple classifiers. Finally, since lower (higher) compression ratio can remove both perturbations and original features slightly (aggressively), we use these trained multiple models for model ensemble. The majority vote of the ensemble of models is adopted as final predictions. Experiments results show our method can improve defense efficiency while maintaining original accuracy.

cs.CV↗

arXiv2021

Meta-Learning Divergences of Variational Inference

Variational inference (VI) plays an essential role in approximate Bayesian inference due to its computational efficiency and broad applicability. Crucial to the performance of VI is the selection of the associated divergence measure, as VI approximates the intractable distribution by minimizing this divergence. In this paper we propose a meta-learning algorithm to learn the divergence metric suited for the task of interest, automating the design of VI methods. In addition, we learn the initialization of the variational parameters without additional cost when our method is deployed in the few-shot learning scenarios. We demonstrate our approach outperforms standard VI on Gaussian mixture distribution approximation, Bayesian neural network regression, image generation with variational autoencoders and recommender systems with a partial variational autoencoder.

cs.LG↗

arXiv2021

On an inverse scattering transform for the nonlinear Schödinger equation on the half-line

In this paper, we develop an inverse scattering transform for the integrable focusing nonlinear Schrödinger (NLS) equation on the half-line subject to a class of boundary conditions. The method is based on the notions of integrable boundary conditions and double-row monodromy matrix, developed by Sklyanin, which characterize initial-boundary value problems for NLS on an interval. It follows from Sklyanin's approach that a hierarchy of integrable boundary conditions can be derived based on a semi-discrete type Lax pair involving the time-part Lax matrix of NLS and a reflection matrix. The inverse scattering transform relies on the formulation of a scattering system for the double-row monodromy matrix characterizing an interval problem. The scattering system for the half-line problem for NLS is obtained by extending one endpoint of the interval to infinity. Then, we derive spectral and analytic properties of the scattering systems, and set up the inverse part using a Riemann-Hilbert formulation. We also show that our approach is equivalent to a nonlinear method of reflection by extending the initial-boundary value problem on the half-line to an initial value on the whole axis. Explicit examples of soliton solutions on the half-line are provided. Although we only consider the NLS model as a particular example, the method we present in this paper can be readily applied to a wide range of integrable PDEs on the half-line.

nlin.SI↗

arXiv2021

Probabilistic DAG Search

Exciting contemporary machine learning problems have recently been phrased in the classic formalism of tree search -- most famously, the game of Go. Interestingly, the state-space underlying these sequential decision-making problems often posses a more general latent structure than can be captured by a tree. In this work, we develop a probabilistic framework to exploit a search space's latent structure and thereby share information across the search tree. The method is based on a combination of approximate inference in jointly Gaussian models for the explored part of the problem, and an abstraction for the unexplored part that imposes a reduction of complexity ad hoc. We empirically find our algorithm to compare favorably to existing non-probabilistic alternatives in Tic-Tac-Toe and a feature selection application.

cs.LG↗

arXiv2021

Causally Constrained Data Synthesis for Private Data Release

Making evidence based decisions requires data. However for real-world applications, the privacy of data is critical. Using synthetic data which reflects certain statistical properties of the original data preserves the privacy of the original data. To this end, prior works utilize differentially private data release mechanisms to provide formal privacy guarantees. However, such mechanisms have unacceptable privacy vs. utility trade-offs. We propose incorporating causal information into the training process to favorably modify the aforementioned trade-off. We theoretically prove that generative models trained with additional causal knowledge provide stronger differential privacy guarantees. Empirically, we evaluate our solution comparing different models based on variational auto-encoders (VAEs), and show that causal information improves resilience to membership inference, with improvements in downstream utility.

cs.LG↗

arXiv2021

Squared eigenfunction symmetry of the D$Δ$mKP hierarchy and its constraint

In this paper squared eigenfunction symmetry of the differential-difference modified Kadomtsev-Petviashvili (D$Δ$mKP) hierarchy and its constraint are considered. Under the constraint, the Lax triplets of the D$Δ$mKP hierarchy, together with their adjoint forms, give rise to the positive relativistic Toda (R-Toda) hierarchy. An invertible transformation is given to connect the positive and negative R-Toda hierarchies. The positive R-Toda hierarchy is reduced to the differential-difference Burgers hierarchy. We also consider another D$Δ$mKP hierarchy and show that its squared eigenfunction symmetry constraint gives rise to the Volterra hierarchy. In addition, we revisit the Ragnisco-Tu hierarchy which is a squared eigenfunction symmetry constraint of the differential-difference Kadomtsev-Petviashvili (D$Δ$KP) system. It was thought the Ragnisco-Tu hierarchy does not exist one-field reduction, but here we find an one-field reduction to reduce the hierarchy to the Volterra hierarchy. Besides, the differential-difference Burgers hierarchy are also investigated in Appendix. A multi-dimensionally consistent 3-point discrete Burgers equation is given.

nlin.SI↗

arXiv2021

Tunable terahertz plasmons in graphite thin films

Tunable terahertz plasmons are essential for reconfigurable photonics, which have been demonstrated in graphene through gating, though with relatively weak responses. Here, we demonstrate strong terahertz plasmons in graphite thin films via infrared spectroscopy, with dramatic tunability by even a moderate temperature change or an in-situ bias voltage. Meanwhile, through magneto-plasmon studies, we reveal that massive electrons and massless Dirac holes make comparable contributions to the plasmon response. Our study not only sets up a platform for further exploration of two-component plasmas, but also opens an avenue for terahertz modulation through electrical bias or all-optical means.

physics.optics↗

arXiv2021

Pressure-enhanced ferromagnetism in layered CrSiTe3 flakes

The research on van der Waals (vdW) layered ferromagnets have promoted the development of nanoscale spintronics and applications. However, low-temperature ferromagnetic properties of these materials greatly hinder their applications. Here, we report pressure-enhanced ferromagnetic behaviours in layered CrSiTe3 flakes revealed by high-pressure magnetic circular dichroism (MCD) measurement. At ambient pressure, CrSiTe3 undergoes a paramagnetic-to-ferromagnetic phase transition at 32.8 K, with a negligible hysteresis loop, indicating a soft ferromagnetic behaviour. Under 4.6 GPa pressure, the soft ferromagnet changes into hard one, signalled by a rectangular hysteretic loop with remnant magnetization at zero field. Interestingly, with further increasing pressure, the coercive field (H_c) dramatically increases from 0.02 T at 4.6 GPa to 0.17 T at 7.8 GPa, and the Curie temperature (T_c^h: the temperature for closing the hysteresis loop) also increases from ~36 K at 4.6 GPa to ~138 K at 7.8 GPa. The influences of pressure on exchange interactions are further investigated by density functional theory calculations, which reveal that the in-plane nearest-neighbor exchange interaction and magneto-crystalline anisotropy increase simultaneously as pressure increases, leading to increased H_c and T_c^h in experiments. The effective interaction between magnetic couplings and external pressure offers new opportunities for both searching room-temperature layered ferromagnets and designing pressure-sensitive magnetic functional devices.

cond-mat.mtrl-sci↗

arXiv2021

Dynamic gesture retrieval: searching videos by human pose sequence

The number of static human poses is limited, it is hard to retrieve the exact videos using one single pose as the clue. However, with a pose sequence or a dynamic gesture as the keyword, retrieving specific videos becomes more feasible. We propose a novel method for querying videos containing a designated sequence of human poses, whereas previous works only designate a single static pose. The proposed method takes continuous 3d human poses from keyword gesture video and video candidates, then converts each pose in individual frames into bone direction descriptors, which describe the direction of each natural connection in articulated pose. A temporal pyramid sliding window is then applied to find matches between designated gesture and video candidates, which ensures that same gestures with different duration can be matched.

cs.CV↗

arXiv2021

A data-driven and model-based accelerated Hamiltonian Monte Carlo method for Bayesian elliptic inverse problems

In this paper, we consider a Bayesian inverse problem modeled by elliptic partial differential equations (PDEs). Specifically, we propose a data-driven and model-based approach to accelerate the Hamiltonian Monte Carlo (HMC) method in solving large-scale Bayesian inverse problems. The key idea is to exploit (model-based) and construct (data-based) the intrinsic approximate low-dimensional structure of the underlying problem which consists of two components - a training component that computes a set of data-driven basis to achieve significant dimension reduction in the solution space, and a fast solving component that computes the solution and its derivatives for a newly sampled elliptic PDE with the constructed data-driven basis. Hence we achieve an effective data and model-based approach for the Bayesian inverse problem and overcome the typical computational bottleneck of HMC - repeated evaluation of the Hamiltonian involving the solution (and its derivatives) modeled by a complex system, a multiscale elliptic PDE in our case. We present numerical examples to demonstrate the accuracy and efficiency of the proposed method.

math.NA↗

arXiv2021

Variational Bayesian Supertrees

Given overlapping subsets of a set of taxa (e.g. species), and posterior distributions on phylogenetic tree topologies for each of these taxon sets, how can we infer a posterior distribution on phylogenetic tree topologies for the entire taxon set? Although the equivalent problem for in the non-Bayesian case has attracted substantial research, the Bayesian case has not attracted the attention it deserves. In this paper we develop a variational Bayes approach to this problem and demonstrate its effectiveness.

q-bio.PE↗

arXiv2021

Instructions and Guide for Diagnostic Questions: The NeurIPS 2020 Education Challenge

Digital technologies are becoming increasingly prevalent in education, enabling personalized, high quality education resources to be accessible by students across the world. Importantly, among these resources are diagnostic questions: the answers that the students give to these questions reveal key information about the specific nature of misconceptions that the students may hold. Analyzing the massive quantities of data stemming from students' interactions with these diagnostic questions can help us more accurately understand the students' learning status and thus allow us to automate learning curriculum recommendations. In this competition, participants will focus on the students' answer records to these multiple-choice diagnostic questions, with the aim of 1) accurately predicting which answers the students provide; 2) accurately predicting which questions have high quality; and 3) determining a personalized sequence of questions for each student that best predicts the student's answers. These tasks closely mimic the goals of a real-world educational platform and are highly representative of the educational challenges faced today. We provide over 20 million examples of students' answers to mathematics questions from Eedi, a leading educational platform which thousands of students interact with daily around the globe. Participants to this competition have a chance to make a lasting, real-world impact on the quality of personalized education for millions of students across the world.

cs.CY↗