arXiv subjects

Yixuan Li

Publications and source records attributed to Yixuan Li.

At least 253 records · Page 14Linked to original sources

arXiv2020

Learning Implicit Credit Assignment for Cooperative Multi-Agent Reinforcement Learning

We present a multi-agent actor-critic method that aims to implicitly address the credit assignment problem under fully cooperative settings. Our key motivation is that credit assignment among agents may not require an explicit formulation as long as (1) the policy gradients derived from a centralized critic carry sufficient information for the decentralized agents to maximize their joint action value through optimal cooperation and (2) a sustained level of exploration is enforced throughout training. Under the centralized training with decentralized execution (CTDE) paradigm, we achieve the former by formulating the centralized critic as a hypernetwork such that a latent state representation is integrated into the policy gradients through its multiplicative association with the stochastic policies; to achieve the latter, we derive a simple technique called adaptive entropy regularization where magnitudes of the entropy gradients are dynamically rescaled based on the current policy stochasticity to encourage consistent levels of exploration. Our algorithm, referred to as LICA, is evaluated on several benchmarks including the multi-agent particle environments and a set of challenging StarCraft II micromanagement tasks, and we show that LICA significantly outperforms previous methods.

cs.LG↗

arXiv2020

Enhancing The Reliability of Out-of-distribution Image Detection in Neural Networks

We consider the problem of detecting out-of-distribution images in neural networks. We propose ODIN, a simple and effective method that does not require any change to a pre-trained neural network. Our method is based on the observation that using temperature scaling and adding small perturbations to the input can separate the softmax score distributions between in- and out-of-distribution images, allowing for more effective detection. We show in a series of experiments that ODIN is compatible with diverse network architectures and datasets. It consistently outperforms the baseline approach by a large margin, establishing a new state-of-the-art performance on this task. For example, ODIN reduces the false positive rate from the baseline 34.7% to 4.3% on the DenseNet (applied to CIFAR-10) when the true positive rate is 95%.

cs.LG↗

arXiv2020

Actions as Moving Points

The existing action tubelet detectors often depend on heuristic anchor design and placement, which might be computationally expensive and sub-optimal for precise localization. In this paper, we present a conceptually simple, computationally efficient, and more precise action tubelet detection framework, termed as MovingCenter Detector (MOC-detector), by treating an action instance as a trajectory of moving points. Based on the insight that movement information could simplify and assist action tubelet detection, our MOC-detector is composed of three crucial head branches: (1) Center Branch for instance center detection and action recognition, (2) Movement Branch for movement estimation at adjacent frames to form trajectories of moving points, (3) Box Branch for spatial extent detection by directly regressing bounding box size at each estimated center. These three branches work together to generate the tubelet detection results, which could be further linked to yield video-level tubes with a matching strategy. Our MOC-detector outperforms the existing state-of-the-art methods for both metrics of frame-mAP and video-mAP on the JHMDB and UCF101-24 datasets. The performance gap is more evident for higher video IoU, demonstrating that our MOC-detector is particularly effective for more precise action detection. We provide the code at https://github.com/MCG-NJU/MOC-Detector.

cs.CV↗

arXiv2020

Model Patching: Closing the Subgroup Performance Gap with Data Augmentation

Classifiers in machine learning are often brittle when deployed. Particularly concerning are models with inconsistent performance on specific subgroups of a class, e.g., exhibiting disparities in skin cancer classification in the presence or absence of a spurious bandage. To mitigate these performance differences, we introduce model patching, a two-stage framework for improving robustness that encourages the model to be invariant to subgroup differences, and focus on class information shared by subgroups. Model patching first models subgroup features within a class and learns semantic transformations between them, and then trains a classifier with data augmentations that deliberately manipulate subgroup features. We instantiate model patching with CAMEL, which (1) uses a CycleGAN to learn the intra-class, inter-subgroup augmentations, and (2) balances subgroup performance using a theoretically-motivated subgroup consistency regularizer, accompanied by a new robust objective. We demonstrate CAMEL's effectiveness on 3 benchmark datasets, with reductions in robust error of up to 33% relative to the best baseline. Lastly, CAMEL successfully patches a model that fails due to spurious features on a real-world skin cancer dataset.

cs.LG↗

arXiv2020

Deep-learning-enabled geometric constraints and phase unwrapping for single-shot absolute 3D shape measurement

Fringe projection profilometry (FPP) is one of the most popular three-dimensional (3D) shape measurement techniques, and has becoming more prevalently adopted in intelligent manufacturing, defect detection and some other important applications. In FPP, how to efficiently recover the absolute phase has always been a great challenge. The stereo phase unwrapping (SPU) technologies based on geometric constraints can eliminate phase ambiguity without projecting any additional fringe patterns, which maximizes the efficiency of the retrieval of absolute phase. Inspired by the recent success of deep learning technologies for phase analysis, we demonstrate that deep learning can be an effective tool that organically unifies the phase retrieval, geometric constraints, and phase unwrapping steps into a comprehensive framework. Driven by extensive training dataset, the neutral network can gradually "learn" how to transfer one high-frequency fringe pattern into the "physically meaningful", and "most likely" absolute phase, instead of "step by step" as in convention approaches. Based on the properly trained framework, high-quality phase retrieval and robust phase ambiguity removal can be achieved based on only single-frame projection. Experimental results demonstrate that compared with traditional SPU, our method can more efficiently and stably unwrap the phase of dense fringe images in a larger measurement volume with fewer camera views. Limitations about the proposed approach are also discussed. We believe the proposed approach represents an important step forward in high-speed, high-accuracy, motion-artifacts-free absolute 3D shape measurement for complicated object from a single fringe pattern.

eess.IV↗

arXiv2019

Defense Against Adversarial Images using Web-Scale Nearest-Neighbor Search

A plethora of recent work has shown that convolutional networks are not robust to adversarial images: images that are created by perturbing a sample from the data distribution as to maximize the loss on the perturbed example. In this work, we hypothesize that adversarial perturbations move the image away from the image manifold in the sense that there exists no physical process that could have produced the adversarial image. This hypothesis suggests that a successful defense mechanism against adversarial images should aim to project the images back onto the image manifold. We study such defense mechanisms, which approximate the projection onto the unknown image manifold by a nearest-neighbor search against a web-scale image database containing tens of billions of images. Empirical evaluations of this defense strategy on ImageNet suggest that it is very effective in attack settings in which the adversary does not have access to the image database. We also propose two novel attack methods to break nearest-neighbor defenses, and demonstrate conditions under which nearest-neighbor defense fails. We perform a series of ablation experiments, which suggest that there is a trade-off between robustness and accuracy in our defenses, that a large image database (with hundreds of millions of images) is crucial to get good performance, and that careful construction the image database is important to be robust against attacks tailored to circumvent our defenses.

cs.CV↗

arXiv2019

Multiphase unsteady Stokes flow over grooved surfaces: an analytical study

This paper studies multiphase flow within grooved textures exposed to external unsteadiness. We derive analytical expressions for multiphase unsteady Stokes flow driven by oscillating streamwise/spanwise velocity in the presence of periodic grooves. Good agreement is obtained between the analytical solution and direct numerical simulations performed using the volume of fluid method. The effects of oscillation frequency, Reynolds number, and the multiphase interface location on the transfer function between the input signal external to the groove and output near the interface, are examined. The effective slip length and the shear stress over the grooved plane are described.

physics.flu-dyn↗

arXiv2019

Cosmological consequences of Omnia Sequestra

Vacuum energy sequestering is a mechanism for cancelling off the radiative corrections to vacuum energy. The most comprehensive low energy model, proposed in 1606.04958 [hep-th], has the ability to cancel off radiative corrections to vacuum energy for loops of both matter and gravitons, in contrast to earlier proposals. Here we explore the framework and cosmological consequences of this model, which we call Omnia Sequestra, owing to its ability to sequester all. A computation of historic integrals on a cosmological background reveals some subtleties with UV sensitivity not seen in earlier models, but which are tamed in a Universe that grows sufficiently old. For these old Universes, we estimate the size of the radiatively stable residual cosmological constant and show that it does not exceed the critical density of our Universe today. We also study the effect of phase transitions, both homogeneous and inhomogeneous, and find that generically spacetime regions with a small cosmological constant do not need to be fine-tuned against the scale of the transition, a result which is now seen to hold across all models of sequestering. The model is developed in other ways, from its compatibility with inflation, to the correct inclusion of boundaries and the geometric consequences of certain choices of boundary data.

hep-th↗

arXiv2018

Enhancing Stock Market Prediction with Extended Coupled Hidden Markov Model over Multi-Sourced Data

Traditional stock market prediction methods commonly only utilize the historical trading data, ignoring the fact that stock market fluctuations can be impacted by various other information sources such as stock related events. Although some recent works propose event-driven prediction approaches by considering the event data, how to leverage the joint impacts of multiple data sources still remains an open research problem. In this work, we study how to explore multiple data sources to improve the performance of the stock prediction. We introduce an Extended Coupled Hidden Markov Model incorporating the news events with the historical trading data. To address the data sparsity issue of news events for each single stock, we further study the fluctuation correlations between the stocks and incorporate the correlations into the model to facilitate the prediction task. Evaluations on China A-share market data in 2016 show the superior performance of our model against previous methods.

q-fin.ST↗

arXiv2018

Exploring the Limits of Weakly Supervised Pretraining

State-of-the-art visual perception models for a wide range of tasks rely on supervised pretraining. ImageNet classification is the de facto pretraining task for these models. Yet, ImageNet is now nearly ten years old and is by modern standards "small". Even so, relatively little is known about the behavior of pretraining with datasets that are multiple orders of magnitude larger. The reasons are obvious: such datasets are difficult to collect and annotate. In this paper, we present a unique study of transfer learning with large convolutional networks trained to predict hashtags on billions of social media images. Our experiments demonstrate that training for large-scale hashtag prediction leads to excellent results. We show improvements on several image classification and object detection tasks, and report the highest ImageNet-1k single-crop, top-1 accuracy to date: 85.4% (97.6% top-5). We also perform extensive experiments that provide novel empirical data on the relationship between large-scale pretraining and transfer learning performance.

cs.CV↗

arXiv2018

Understanding the Loss Surface of Neural Networks for Binary Classification

It is widely conjectured that the reason that training algorithms for neural networks are successful because all local minima lead to similar performance, for example, see (LeCun et al., 2015, Choromanska et al., 2015, Dauphin et al., 2014). Performance is typically measured in terms of two metrics: training performance and generalization performance. Here we focus on the training performance of single-layered neural networks for binary classification, and provide conditions under which the training error is zero at all local minima of a smooth hinge loss function. Our conditions are roughly in the following form: the neurons have to be strictly convex and the surrogate loss function should be a smooth version of hinge loss. We also provide counterexamples to show that when the loss function is replaced with quadratic loss or logistic loss, the result may not hold.

cs.LG↗

arXiv2017

Stacked Generative Adversarial Networks

In this paper, we propose a novel generative model named Stacked Generative Adversarial Networks (SGAN), which is trained to invert the hierarchical representations of a bottom-up discriminative network. Our model consists of a top-down stack of GANs, each learned to generate lower-level representations conditioned on higher-level representations. A representation discriminator is introduced at each feature hierarchy to encourage the representation manifold of the generator to align with that of the bottom-up discriminative network, leveraging the powerful discriminative representations to guide the generative model. In addition, we introduce a conditional loss that encourages the use of conditional information from the layer above, and a novel entropy loss that maximizes a variational lower bound on the conditional entropy of generator outputs. We first train each stack independently, and then train the whole model end-to-end. Unlike the original GAN that uses a single noise vector to represent all the variations, our SGAN decomposes variations into multiple levels and gradually resolves uncertainties in the top-down generative process. Based on visual inspection, Inception scores and visual Turing test, we demonstrate that SGAN is able to generate images of much higher quality than GANs without stacking.

cs.CV↗

arXiv2017

Snapshot Ensembles: Train 1, get M for free

Ensembles of neural networks are known to be much more robust and accurate than individual networks. However, training multiple deep networks for model averaging is computationally expensive. In this paper, we propose a method to obtain the seemingly contradictory goal of ensembling multiple neural networks at no additional training cost. We achieve this goal by training a single neural network, converging to several local minima along its optimization path and saving the model parameters. To obtain repeated rapid convergence, we leverage recent work on cyclic learning rate schedules. The resulting technique, which we refer to as Snapshot Ensembling, is simple, yet surprisingly effective. We show in a series of experiments that our approach is compatible with diverse network architectures and learning tasks. It consistently yields lower error rates than state-of-the-art single models at no additional training cost, and compares favorably with traditional network ensembles. On CIFAR-10 and CIFAR-100 our DenseNet Snapshot Ensembles obtain error rates of 3.4% and 17.4% respectively.

cs.LG↗

arXiv2016

Scalable and Robust Local Community Detection via Adaptive Subgraph Extraction and Diffusions

Local community detection, the problem of identifying a set of relevant nodes nearby a small set of input seed nodes, is an important graph primitive with a wealth of applications and research activity. Recent approaches include using local spectral information, graph diffusions, and random walks to determine a community from input seeds. As networks grow to billions of nodes and exhibit diverse structures, it is important that community detection algorithms are not only efficient, but also robust to different structural features. Toward this goal, we explore pre-processing techniques and modifications to existing local methods aimed at improving the scalability and robustness of algorithms related to community detection. Experiments show that our modifications improve both speed and quality of existing methods for locating ground truth communities, and are more robust across graphs and communities of varying sizes, densities, and diameters. Our subgraph extraction method uses adaptively selected PageRank parameters to improve on the recall and runtime of a walk-based pre-processing technique of Li et al. for extracting subgraphs before searching for a community. We then use this technique to enable the first scalable implementation of the recent Local Fiedler method of Mahoney et al. Our experimental evaluation shows our pre-processed version of Local Fiedler, as well as our novel simplification of the LEMON community detection framework of Li et al., offer significant speedups over their predecessors and obtain cluster quality competitive with the state of the art.

cs.SI↗

arXiv2016

Deep Manifold Traversal: Changing Labels with Convolutional Features

Many tasks in computer vision can be cast as a "label changing" problem, where the goal is to make a semantic change to the appearance of an image or some subject in an image in order to alter the class membership. Although successful task-specific methods have been developed for some label changing applications, to date no general purpose method exists. Motivated by this we propose deep manifold traversal, a method that addresses the problem in its most general form: it first approximates the manifold of natural images then morphs a test image along a traversal path away from a source class and towards a target class while staying near the manifold throughout. The resulting algorithm is surprisingly effective and versatile. It is completely data driven, requiring only an example set of images from the desired source and target domains. We demonstrate deep manifold traversal on highly diverse label changing tasks: changing an individual's appearance (age and hair color), changing the season of an outdoor image, and transforming a city skyline towards nighttime.

cs.LG↗

arXiv2016

Convergent Learning: Do different neural networks learn the same representations?

Recent success in training deep neural networks have prompted active investigation into the features learned on their intermediate layers. Such research is difficult because it requires making sense of non-linear computations performed by millions of parameters, but valuable because it increases our ability to understand current models and create improved versions of them. In this paper we investigate the extent to which neural networks exhibit what we call convergent learning, which is when the representations learned by multiple nets converge to a set of features which are either individually similar between networks or where subsets of features span similar low-dimensional spaces. We propose a specific method of probing representations: training multiple networks and then comparing and contrasting their individual, learned representations at the level of neurons or groups of neurons. We begin research into this question using three techniques to approximately align different neural networks on a feature level: a bipartite matching approach that makes one-to-one assignments between neurons, a sparse prediction approach that finds one-to-many mappings, and a spectral clustering approach that finds many-to-many mappings. This initial investigation reveals a few previously unknown properties of neural networks, and we argue that future research into the question of convergent learning will yield many more. The insights described here include (1) that some features are learned reliably in multiple networks, yet other features are not consistently learned; (2) that units learn to span low-dimensional subspaces and, while these subspaces are common to multiple networks, the specific basis vectors learned are not; (3) that the representation codes show evidence of being a mix between a local code and slightly, but not fully, distributed codes across multiple units.

cs.LG↗

arXiv2016

The Lifecycle and Cascade of WeChat Social Messaging Groups

Social instant messaging services are emerging as a transformative form with which people connect, communicate with friends in their daily life - they catalyze the formation of social groups, and they bring people stronger sense of community and connection. However, research community still knows little about the formation and evolution of groups in the context of social messaging - their lifecycles, the change in their underlying structures over time, and the diffusion processes by which they develop new members. In this paper, we analyze the daily usage logs from WeChat group messaging platform - the largest standalone messaging communication service in China - with the goal of understanding the processes by which social messaging groups come together, grow new members, and evolve over time. Specifically, we discover a strong dichotomy among groups in terms of their lifecycle, and develop a separability model by taking into account a broad range of group-level features, showing that long-term and short-term groups are inherently distinct. We also found that the lifecycle of messaging groups is largely dependent on their social roles and functions in users' daily social experiences and specific purposes. Given the strong separability between the long-term and short-term groups, we further address the problem concerning the early prediction of successful communities. In addition to modeling the growth and evolution from group-level perspective, we investigate the individual-level attributes of group members and study the diffusion process by which groups gain new members. By considering members' historical engagement behavior as well as the local social network structure that they embedded in, we develop a membership cascade model and demonstrate the effectiveness by achieving AUC of 95.31% in predicting inviter, and an AUC of 98.66% in predicting invitee.

cs.SI↗

arXiv2016

In a World That Counts: Clustering and Detecting Fake Social Engagement at Scale

How can web services that depend on user generated content discern fake social engagement activities by spammers from legitimate ones? In this paper, we focus on the social site of YouTube and the problem of identifying bad actors posting inorganic contents and inflating the count of social engagement metrics. We propose an effective method, Leas (Local Expansion at Scale), and show how the fake engagement activities on YouTube can be tracked over time by analyzing the temporal graph based on the engagement behavior pattern between users and YouTube videos. With the domain knowledge of spammer seeds, we formulate and tackle the problem in a semi-supervised manner --- with the objective of searching for individuals that have similar pattern of behavior as the known seeds --- based on a graph diffusion process via local spectral subspace. We offer a fast, scalable MapReduce deployment adapted from the localized spectral clustering algorithm. We demonstrate the effectiveness of our deployment at Google by achieving an manual review accuracy of 98% on YouTube Comments graph in practice. Comparing with the state-of-the-art algorithm CopyCatch, Leas achieves 10 times faster running time. Leas is actively in use at Google, searching for daily deceptive practices on YouTube's engagement graph spanning over a billion users.

cs.SI↗