Neural language representation models such as BERT pre-trained on large-scale corpora can well capture rich semantic patterns from plain text, and be fine-tuned to consistently improve the performance of various NLP tasks. However, the existing pre-trained language models rarely consider incorporating knowledge graphs (KGs), which can provide rich structured knowledge facts for better language understanding. We argue that informative entities in KGs can enhance language representation with external knowledge. In this paper, we utilize both large-scale textual corpora and KGs to train an enhanced language representation model (ERNIE), which can take full advantage of lexical, syntactic, and knowledge information simultaneously. The experimental results have demonstrated that ERNIE achieves significant improvements on various knowledge-driven tasks, and meanwhile is comparable with the state-of-the-art model BERT on other common NLP tasks. The code and datasets will be available in the future.

ERNIE: Enhanced Language Representation with Informative Entities

Learning from data has led to paradigm shifts in a multitude of disciplines, including web, text and image search, speech recognition, as well as bioinformatics. Can machine learning enable similar breakthroughs in understanding quantum many-body systems? Here we develop an efficient deep learning approach that enables spatially and chemically resolved insights into quantum-mechanical observables of molecular systems. We unify concepts from many-body Hamiltonians with purpose-designed deep tensor neural networks, which leads to size-extensive and uniformly accurate (1 kcal mol−1) predictions in compositional and configurational chemical space for molecules of intermediate size. As an example of chemical relevance, the model reveals a classification of aromatic rings with respect to their stability. Further applications of our model for predicting atomic energies and local chemical potentials in molecules, reliable isomer energies, and molecules with peculiar electronic structure demonstrate the potential of machine learning for revealing insights into complex quantum-chemical systems.

/pdf/quantum-chemical-insights-from-deep-tensor-neural-networks-266u885pa4.pdf

Quantum-Chemical Insights from Deep Tensor Neural Networks

We address the problem of adversarial attacks on text classification, which is rarely studied comparing to attacks on image classification. The challenge of this task is to generate adversarial examples that maintain lexical correctness, grammatical correctness and semantic similarity. Based on the synonyms substitution strategy, we introduce a new word replacement order determined by both the word saliency and the classification probability, and propose a greedy algorithm called probability weighted word saliency (PWWS) for text adversarial attack. Experiments on three popular datasets using convolutional as well as LSTM models show that PWWS reduces the classification accuracy to the most extent, and keeps a very low word substitution rate. A human evaluation study shows that our generated adversarial examples maintain the semantic similarity well and are hard for humans to perceive. Performing adversarial training using our perturbed datasets improves the robustness of the models. At last, our method also exhibits a good transferability on the generated adversarial examples.

/pdf/generating-natural-language-adversarial-examples-through-1ftt7y125j.pdf

Generating Natural Language Adversarial Examples through Probability Weighted Word Saliency.

Deep learning--based models have surpassed classical machine learning--based approaches in various text classification tasks, including sentiment analysis, news categorization, question answering, and natural language inference. In this article, we provide a comprehensive review of more than 150 deep learning--based models for text classification developed in recent years, and we discuss their technical contributions, similarities, and strengths. We also provide a summary of more than 40 popular datasets widely used for text classification. Finally, we provide a quantitative analysis of the performance of different deep learning models on popular benchmarks, and we discuss future research directions.

Deep Learning--based Text Classification: A Comprehensive Review

The combination of modern scientific computing with electronic structure theory can lead to an unprecedented amount of data amenable to intelligent data analysis for the identification of meaningful, novel, and predictive structure-property relationships. Such relationships enable high-throughput screening for relevant properties in an exponentially growing pool of virtual compounds that are synthetically accessible. Here, we present a machine learning (ML) model, trained on a data base of \textit{ab initio} calculation results for thousands of organic molecules, that simultaneously predicts multiple electronic ground- and excited-state properties. The properties include atomization energy, polarizability, frontier orbital eigenvalues, ionization potential, electron affinity, and excitation energies. The ML model is based on a deep multi-task artificial neural network, exploiting underlying correlations between various molecular properties. The input is identical to \emph{ab initio} methods, \emph{i.e.} nuclear charges and Cartesian coordinates of all atoms. For small organic molecules the accuracy of such a "Quantum Machine" is similar, and sometimes superior, to modern quantum-chemical methods---at negligible computational cost.

Machine Learning of Molecular Electronic Properties in Chemical Compound Space

We present Segment-level Neural CRF, which combines neural networks with a linear chain CRF for segment-level sequence modeling tasks such as named entity recognition (NER) and syntactic chunking. Our segment-level CRF can consider higher-order label dependencies compared with conventional word-level CRF. Since it is difficult to consider all possible variable length segments, our method uses segment lattice constructed from the word-level tagging model to reduce the search space. Performing experiments on NER and chunking, we demonstrate that our method outperforms conventional word-level CRF with neural networks.

Segment-Level Neural Conditional Random Fields for Named Entity Recognition

Recently, optical proximity correction model calibration techniques that use SEM contours have enabled possibly
significant improvements in complex mask design. However, compared to conventional CD-based calibration, contour-based
calibration results in increased errors in 1D features. In fact, our research shows that there is a ~1-nm gap, which
we call "CD-gap," between CD measurements directly calculated from a SEM image and CD measurements calculated
from SEM contours. To achieve accurate calibration, SEM contours must match the corresponding CD measurements.
We have developed a CD-gap-free contour extraction technique in response to this problem. In our technique, a mask
edge is classified into shape structures and an optimized SEM contour extraction method is prepared for each shape
structure to reduce the CD-gap. Experimental results show that the CD-gap can be decreased to sub-nm, which clearly
demonstrates the potential of our proposed technique to play a vital role in the lithography process.

A CD-gap-free contour extraction technique for OPC model calibration

We propose a new length-controllable abstractive summarization model. Recent state-of-the-art abstractive summarization models based on encoder-decoder models generate only one summary per source text. However, controllable summarization, especially of the length, is an important aspect for practical applications. Previous studies on length-controllable abstractive summarization incorporate length embeddings in the decoder module for controlling the summary length. Although the length embeddings can control where to stop decoding, they do not decide which information should be included in the summary within the length constraint. Unlike the previous models, our length-controllable abstractive summarization model incorporates a word-level extractive module in the encoder-decoder model instead of length embeddings. Our model generates a summary in two steps. First, our word-level extractor extracts a sequence of important words (we call it the "prototype text") from the source text according to the word-level importance scores and the length constraint. Second, the prototype text is used as additional input to the encoder-decoder model, which generates a summary by jointly encoding and copying words from both the prototype text and source text. Since the prototype text is a guide to both the content and length of the summary, our model can generate an informative and length-controlled summary. Experiments with the CNN/Daily Mail dataset and the NEWSROOM dataset show that our model outperformed previous models in length-controlled settings.

Length-controllable Abstractive Summarization by Guiding with Summary Prototype

Existing methods for Japanese predicate argument structure (PAS) analysis identify case arguments of each predicate without considering interactions between the target PAS and others in a sentence. However, the argument structures of the predicates in a sentence are semantically related to each other. This paper proposes new methods for Japanese PAS analysis to jointly identify case arguments of all predicates in a sentence by (1) modeling multiple PAS interactions with a bipartite graph and (2) approximately searching optimal PAS combinations. Performing experiments on the NAIST Text Corpus, we demonstrate that our joint analysis methods substantially outperform a strong baseline and are comparable to previous work.

/pdf/joint-case-argument-identification-for-japanese-predicate-5akllqp59h.pdf

Joint Case Argument Identification for Japanese Predicate Argument Structure Analysis

Understanding procedural text requires tracking entities, actions and effects as the narrative unfolds. We focus on the challenging real-world problem of action-graph extraction from materials science papers, where language is highly specialized and data annotation is expensive and scarce. We propose a novel approach, Text2Quest, where procedural text is interpreted as instructions for an interactive game. A learning agent completes the game by executing the procedure correctly in a text-based simulated lab environment. The framework can complement existing approaches and enables richer forms of learning compared to static texts. We discuss potential limitations and advantages of the approach, and release a prototype proof-of-concept, hoping to encourage research in this direction.

Hiroyuki Shindo

Papers

Segment-Level Neural Conditional Random Fields for Named Entity Recognition

A CD-gap-free contour extraction technique for OPC model calibration

Length-controllable Abstractive Summarization by Guiding with Summary Prototype

Joint Case Argument Identification for Japanese Predicate Argument Structure Analysis

Playing by the Book: An Interactive Game Approach for Action Graph Extraction from Text