[신간의 별자리x] 우리/미술, 그리고 ‘슬픔의 박물관’

TensorFlow is a machine learning system that operates at large scale and in heterogeneous environments. Tensor-Flow uses dataflow graphs to represent computation, shared state, and the operations that mutate that state. It maps the nodes of a dataflow graph across many machines in a cluster, and within a machine across multiple computational devices, including multicore CPUs, general-purpose GPUs, and custom-designed ASICs known as Tensor Processing Units (TPUs). This architecture gives flexibility to the application developer: whereas in previous "parameter server" designs the management of shared state is built into the system, TensorFlow enables developers to experiment with novel optimizations and training algorithms. TensorFlow supports a variety of applications, with a focus on training and inference on deep neural networks. Several Google services use TensorFlow in production, we have released it as an open-source project, and it has become widely used for machine learning research. In this paper, we describe the TensorFlow dataflow model and demonstrate the compelling performance that TensorFlow achieves for several real-world applications.

/pdf/tensorflow-a-system-for-large-scale-machine-learning-10wgdjhl6k.pdf

TensorFlow: a system for large-scale machine learning

TensorFlow is an interface for expressing machine learning algorithms, and an
implementation for executing such algorithms. A computation expressed using
TensorFlow can be executed with little or no change on a wide variety of
heterogeneous systems, ranging from mobile devices such as phones and tablets
up to large-scale distributed systems of hundreds of machines and thousands of
computational devices such as GPU cards. The system is flexible and can be used
to express a wide variety of algorithms, including training and inference
algorithms for deep neural network models, and it has been used for conducting
research and for deploying machine learning systems into production across more
than a dozen areas of computer science and other fields, including speech
recognition, computer vision, robotics, information retrieval, natural language
processing, geographic information extraction, and computational drug
discovery. This paper describes the TensorFlow interface and an implementation
of that interface that we have built at Google. The TensorFlow API and a
reference implementation were released as an open-source package under the
Apache 2.0 license in November, 2015 and are available at www.tensorflow.org.

/pdf/tensorflow-large-scale-machine-learning-on-heterogeneous-2yk4fgchw6.pdf

TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems

How to Do Things With Words

A goal of statistical language modeling is to learn the joint probability function of sequences of words in a language. This is intrinsically difficult because of the curse of dimensionality: a word sequence on which the model will be tested is likely to be different from all the word sequences seen during training. Traditional but very successful approaches based on n-grams obtain generalization by concatenating very short overlapping sequences seen in the training set. We propose to fight the curse of dimensionality by learning a distributed representation for words which allows each training sentence to inform the model about an exponential number of semantically neighboring sentences. The model learns simultaneously (1) a distributed representation for each word along with (2) the probability function for word sequences, expressed in terms of these representations. Generalization is obtained because a sequence of words that has never been seen before gets high probability if it is made of words that are similar (in the sense of having a nearby representation) to words forming an already seen sentence. Training such large models (with millions of parameters) within a reasonable time is itself a significant challenge. We report on experiments using neural networks for the probability function, showing on two text corpora that the proposed approach significantly improves on state-of-the-art n-gram models, and that the proposed approach allows to take advantage of longer contexts.

/pdf/a-neural-probabilistic-language-model-v51mwoff5f.pdf

A neural probabilistic language model

In this paper we explore a new theory of discourse structure that stresses the role of purpose and processing in discourse. In this theory, discourse structure is composed of three separate but interrelated components: the structure of the sequence of utterances (called the linguistic structure), a structure of purposes (called the intentional structure), and the state of focus of attention (called the attentional state). The linguistic structure consists of segments of the discourse into which the utterances naturally aggregate. The intentional structure captures the discourse-relevant purposes, expressed in each of the linguistic segments as well as relationships among them. The attentional state is an abstraction of the focus of attention of the participants as the discourse unfolds. The attentional state, being dynamic, records the objects, properties, and relations that are salient at each point of the discourse. The distinction among these components is essential to provide an adequate explanation of such discourse phenomena as cue phrases, referring expressions, and interruptions.The theory of attention, intention, and aggregation of utterances is illustrated in the paper with a number of example discourses. Various properties of discourse are described, and explanations for the behavior of cue phrases, referring expressions, and interruptions are explored.This theory provides a framework for describing the processing of utterances in a discourse. Discourse processing requires recognizing how the utterances of the discourse aggregate into segments, recognizing the intentions expressed in the discourse and the relationships among intentions, and tracking the discourse through the operation of the mechanisms associated with attentional state. This processing description specifies in these recognition tasks the role of information from the discourse and from the participants' knowledge of the domain.

/pdf/attention-intentions-and-the-structure-of-discourse-sat0ew257g.pdf

Attention, intentions, and the structure of discourse

This paper concerns relationships among focus of attention, choice of referring expression, and perceived coherence of utterances within a discourse segment. It presents a framework and initial theory of centering intended to model the local component of attentional state. The paper examines interactions between local coherence and choice of referring expressions; it argues that differences in coherence correspond in part to the inference demands made by different types of referring expressions, given a particular attentional state. It demonstrates that the attentional state properties modeled by centering can account for these differences.

/pdf/centering-a-framework-for-modeling-the-local-coherence-of-hvb4n7dktx.pdf

Centering: a framework for modeling the local coherence of discourse

https://dash.harvard.edu/bitstream/handle/1/2562070/Grosz_CollaborativePlans.pdf?sequence=2

Collaborative plans for complex group action

Natural language processing

https://www.onlinelibrary.wiley.com/doi/pdf/10.1207/s15516709cog1703_1

Barbara J. Grosz

Papers

Attention, intentions, and the structure of discourse

Centering: a framework for modeling the local coherence of discourse

Collaborative plans for complex group action

Natural language processing

Pronouns, Names, and the Centering of Attention in Discourse