arXiv subjects

Jiaming Song

Publications and source records attributed to Jiaming Song.

82 records · Page 5Linked to original sources

arXiv2018

An Empirical Analysis of Proximal Policy Optimization with Kronecker-factored Natural Gradients

In this technical report, we consider an approach that combines the PPO objective and K-FAC natural gradient optimization, for which we call PPOKFAC. We perform a range of empirical analysis on various aspects of the algorithm, such as sample complexity, training speed, and sensitivity to batch size and training epochs. We observe that PPOKFAC is able to outperform PPO in terms of sample complexity and speed in a range of MuJoCo environments, while being scalable in terms of batch size. In spite of this, it seems that adding more epochs is not necessarily helpful for sample efficiency, and PPOKFAC seems to be worse than its A2C counterpart, ACKTR.

cs.AI↗

arXiv2018

Deep analytics of atomically-resolved images: manifest and latent features

Recent advances in scanning transmission electron and scanning tunneling microscopies allow researchers to measure materials structural and electronic properties, such as atomic displacements and charge density modulations, at an Angstrom scale in real space. At the same time, the ability to quickly acquire large, high-resolution datasets has created a challenge for rapid physics-based analysis of images that typically contain several hundreds to several thousand atomic units. Here we demonstrate a universal deep-learning based framework for locating and characterizing atomic species in the lattice, which can be applied to different types of atomically resolved measurements on different materials. Specifically, by inspecting and categorizing features in the output layer of a convolutional neural network, we are able to detect structural and electronic 'anomalies' associated with the presence of point defects in a tungsten disulfide monolayer, non-uniformity of the charge density distribution around specific lattice sites on the surface of strongly correlated oxides, and transition between different structural states of buckybowl molecules. We further extended our method towards tracking, from one image frame to another, minute distortions in the geometric shape of individual Si dumbbells in a 3-dimensional Si sample, which are associated with a motion of lattice defects and impurities. Due the applicability of our framework to both scanning tunneling microscopy and scanning transmission electron microscopy measurements, it can provide a fast and straightforward way towards creating a unified database of defect-property relationships from experimental data for each material.

cond-mat.mtrl-sci↗

arXiv2017

Direct atomic fabrication and dopant positioning in Si using electron beams with active real time image-based feedback

Semiconductor fabrication is a mainstay of modern civilization, enabling the myriad applications and technologies that underpin everyday life. However, while sub-10 nanometer devices are already entering the mainstream, the end of the Moore's Law roadmap still lacks tools capable of bulk semiconductor fabrication on sub-nanometer and atomic levels, with probe-based manipulation being explored as the only known pathway. Here we demonstrate that the atomic-sized focused beam of a scanning transmission electron microscope can be used to manipulate semiconductors such as Si on the atomic level, inducing growth of crystalline Si from the amorphous phase, reentrant amorphization, milling, and dopant-front motion. These phenomena are visualized in real time with atomic resolution. We further implement active feedback control based on real-time image analytics to control the e-beam motion, enabling shape control and providing a pathway for atom-by-atom correction of fabricated structures in the near future. These observations open a new epoch for atom-by-atom manufacturing in bulk, the long-held dream of nanotechnology.

cond-mat.mtrl-sci↗

arXiv2017

InfoGAIL: Interpretable Imitation Learning from Visual Demonstrations

The goal of imitation learning is to mimic expert behavior without access to an explicit reward signal. Expert demonstrations provided by humans, however, often show significant variability due to latent factors that are typically not explicitly modeled. In this paper, we propose a new algorithm that can infer the latent structure of expert demonstrations in an unsupervised way. Our method, built on top of Generative Adversarial Imitation Learning, can not only imitate complex behaviors, but also learn interpretable and meaningful representations of complex behavioral data, including visual demonstrations. In the driving domain, we show that a model learned from human demonstrations is able to both accurately reproduce a variety of behaviors and accurately anticipate human actions using raw visual inputs. Compared with various baselines, our method can better capture the latent structure underlying expert demonstrations, often recovering semantically meaningful factors of variation in the data.

cs.LG↗

arXiv2017

Learning Hierarchical Features from Generative Models

Deep neural networks have been shown to be very successful at learning feature hierarchies in supervised learning tasks. Generative models, on the other hand, have benefited less from hierarchical models with multiple layers of latent variables. In this paper, we prove that hierarchical latent variable models do not take advantage of the hierarchical structure when trained with existing variational methods, and provide some limitations on the kind of features existing models can learn. Finally we propose an alternative architecture that do not suffer from these limitations. Our model is able to learn highly interpretable and disentangled hierarchical features on several natural image datasets with no task specific regularization or prior knowledge.

cs.LG↗

arXiv2017

On the Limits of Learning Representations with Label-Based Supervision

Advances in neural network based classifiers have transformed automatic feature learning from a pipe dream of stronger AI to a routine and expected property of practical systems. Since the emergence of AlexNet every winning submission of the ImageNet challenge has employed end-to-end representation learning, and due to the utility of good representations for transfer learning, representation learning has become as an important and distinct task from supervised learning. At present, this distinction is inconsequential, as supervised methods are state-of-the-art in learning transferable representations. But recent work has shown that generative models can also be powerful agents of representation learning. Will the representations learned from these generative methods ever rival the quality of those from their supervised competitors? In this work, we argue in the affirmative, that from an information theoretic perspective, generative models have greater potential for representation learning. Based on several experimentally validated assumptions, we show that supervised learning is upper bounded in its capacity for representation learning in ways that certain generative models, such as Generative Adversarial Networks (GANs) are not. We hope that our analysis will provide a rigorous motivation for further exploration of generative representation learning.

cs.LG↗

arXiv2017

Towards Deeper Understanding of Variational Autoencoding Models

We propose a new family of optimization criteria for variational auto-encoding models, generalizing the standard evidence lower bound. We provide conditions under which they recover the data distribution and learn latent features, and formally show that common issues such as blurry samples and uninformative latent features arise when these conditions are not met. Based on these new insights, we propose a new sequential VAE model that can generate sharp samples on the LSUN image dataset based on pixel-wise reconstruction loss, and propose an optimization criterion that encourages unsupervised learning of informative latent features.

cs.LG↗

arXiv2016

Factored Temporal Sigmoid Belief Networks for Sequence Learning

Deep conditional generative models are developed to simultaneously learn the temporal dependencies of multiple sequences. The model is designed by introducing a three-way weight tensor to capture the multiplicative interactions between side information and sequences. The proposed model builds on the Temporal Sigmoid Belief Network (TSBN), a sequential stack of Sigmoid Belief Networks (SBNs). The transition matrices are further factored to reduce the number of parameters and improve generalization. When side information is not available, a general framework for semi-supervised learning based on the proposed model is constituted, allowing robust sequence classification. Experimental results show that the proposed approach achieves state-of-the-art predictive and classification performance on sequential data, and has the capacity to synthesize sequences, with controlled style transitioning and blending.

stat.ML↗

arXiv2016

Max-Margin Nonparametric Latent Feature Models for Link Prediction

Link prediction is a fundamental task in statistical network analysis. Recent advances have been made on learning flexible nonparametric Bayesian latent feature models for link prediction. In this paper, we present a max-margin learning method for such nonparametric latent feature relational models. Our approach attempts to unite the ideas of max-margin learning and Bayesian nonparametrics to discover discriminative latent features for link prediction. It inherits the advances of nonparametric Bayesian methods to infer the unknown latent social dimension, while for discriminative link prediction, it adopts the max-margin learning principle by minimizing a hinge-loss using the linear expectation operator, without dealing with a highly nonlinear link likelihood function. For posterior inference, we develop an efficient stochastic variational inference algorithm under a truncated mean-field assumption. Our methods can scale up to large-scale real networks with millions of entities and tens of millions of positive links. We also provide a full Bayesian formulation, which can avoid tuning regularization hyper-parameters. Experimental results on a diverse range of real datasets demonstrate the benefits inherited from max-margin learning and Bayesian nonparametric inference.

cs.LG↗

arXiv2015

Discriminative Nonparametric Latent Feature Relational Models with Data Augmentation

We present a discriminative nonparametric latent feature relational model (LFRM) for link prediction to automatically infer the dimensionality of latent features. Under the generic RegBayes (regularized Bayesian inference) framework, we handily incorporate the prediction loss with probabilistic inference of a Bayesian model; set distinct regularization parameters for different types of links to handle the imbalance issue in real networks; and unify the analysis of both the smooth logistic log-loss and the piecewise linear hinge loss. For the nonconjugate posterior inference, we present a simple Gibbs sampler via data augmentation, without making restricting assumptions as done in variational methods. We further develop an approximate sampler using stochastic gradient Langevin dynamics to handle large networks with hundreds of thousands of entities and millions of links, orders of magnitude larger than what existing LFRM models can process. Extensive studies on various real networks show promising performance.

cs.LG↗