To parse images into fine-grained semantic parts, the complex fine-grained elements will put it in trouble when using off-the-shelf semantic segmentation networks. In this paper, for image parsing task, we propose to parse images from coarse to fine with progressively refined semantic classes. It is achieved by stacking the segmentation layers in a segmentation network several times. The former segmentation module parses images at a coarser-grained level, and the result will be feed to the following one to provide effective contextual clues for the finer-grained parsing. To recover the details of small structures, we add skip connections from shallow layers of the network to fine-grained parsing modules. As for the network training, we merge classes in groundtruth to get coarse-to-fine label maps, and train the stacked network with these hierarchical supervision end-to-end. Our coarse-to-fine stacked framework can be injected into many advanced neural networks to improve the parsing results. Extensive evaluations on several public datasets including face parsing and human parsing well demonstrate the superiority of our method.

Semantic parsers translate language utterances to programs, but are often trained from utterance-denotation pairs only. Consequently, parsers must overcome the problem of spuriousness at training time, where an incorrect program found at search time accidentally leads to a correct denotation. We propose that in small well-typed domains, we can semi-automatically generate an abstract representation for examples that facilitates information sharing across examples. This alleviates spuriousness, as the probability of randomly obtaining a correct answer from a program decreases across multiple examples. We test our approach on CNLVR, a challenging visual reasoning dataset, where spuriousness is central because denotations are either TRUE or FALSE, and thus random programs have high probability of leading to a correct denotation. We develop the first semantic parser for this task and reach 83.5% accuracy, a 15.7% absolute accuracy improvement compared to the best reported accuracy so far.

Building a semantic parser quickly in a new domain is a fundamental challenge for conversational interfaces, as current semantic parsers require expensive supervision and lack the ability to generalize to new domains. In this paper, we introduce a zero-shot approach to semantic parsing that can parse utterances in unseen domains while only being trained on examples in other source domains. First, we map an utterance to an abstract, domain-independent, logical form that represents the structure of the logical form, but contains slots instead of KB constants. Then, we replace slots with KB constants via lexical alignment scores and global inference. Our model reaches an average accuracy of 53.1% on 7 domains in the Overnight dataset, substantially better than other zero-shot baselines, and performs as good as a parser trained on over 30% of the target domain examples.

However, this variation is often aligned with author attributes such as age, gender, and geography, as well as more readily-available social network metadata. In this paper, we report new evidence on the link between language and social networks in the task of part-of-speech tagging. We find that tagger error rates are correlated with network structure, with high accuracy in some parts of the network, and lower accuracy elsewhere. As a result, tagger accuracy depends on training from a balanced sample of the network, rather than training on texts from a narrow subcommunity. We also describe our attempts to add robustness to stylistic variation, by building a mixture-of-experts model in which each expert is associated with a region of the social network. While prior work found that similar approaches yield performance improvements in sentiment analysis and entity linking, we were unable to obtain performance improvements in part-of-speech tagging, despite strong evidence for the link between part-of-speech error rates and social network structure.

Recent work in natural language processing has looked at learning text to code translation models using parallel pairs of text and code samples from example source code libraries (for a review, see Neubig (2016)). In particular, Richardson and Kuhn (2017a,b); Richardson et al. (2018) look at learning to translate short text descriptions to function signature representations as a first step towards modeling the semantics of function documentation. Examples pairs of docstring and function signature representations are shown in Figure 1; using such pairs, the goal is to learn a general model that can robustly translate a given description of a function to a formal representation of that function. Initially, these datasets were proposed as a synthetic resource for studying semantic parser induction (Mooney, 2007), or for building models that learn to translate text to formal meaning representations from parallel data (see Richardson et al. (2017) for a proposal on using these datasets for the inverse problem of data-to-text generation). To date, we have built around 45 API datasets across 11 popular programming languages (e.g., Python, Java, C, Scheme, Haskell, PHP) and 7 natural languages (see Richardson (2017)), each using an ad hoc rendering of the target function signature representations. In this brief note, we define a unified syntax for expressing these representations, as well as a systematic mapping into first-order logic and a small subject domain model. In doing this, we aim to answer the following question: what do these function signatures that are being learned actually mean, and how can they be used for solving more complex natural language understanding problems (for a similar idea, see Bos (2016))? By recasting the learned representations in terms of classical logic, the hope is that our datasets will in particular be made more accessible to studies on natural language based program synthesis (Raza et al., 2015) and natural language programming more generally. In what follows, we first define a general syntax for these representations, then discuss the mapping into logic and the various applications that motivate our particular approach and subject domain model.

Evaluation in Monolingual Translation, and particularly in Grammatical Error Correction (GEC) is a challenging research field, much due to the difficulty in integrating different types of rewriting operations into a single measure, and the vast number of valid outputs (Tetreault and Chodorow, 2008; Madnani et al., 2011; Chodorow et al., 2012; Bryant and Ng, 2015). These difficulties have recently motivated a number of proposals for new, improved reference-based measures (RBMs) (Dahlmeier and Ng, 2012; Felice and Briscoe, 2015; Napoles et al., 2015). Nevertheless, the size and heterogeneity of the space of valid outputs per sentence often prohibits obtaining a reference set that covers this space well, thereby limiting the applicability of RBMs (Bryant and Ng, 2015).

Previously referred to as `miraculous' in the scientific literature because of its powerful properties and its wide application as optimal solution to the problem of induction/inference, (approximations to) Algorithmic Probability (AP) and the associated Universal Distribution are (or should be) of the greatest importance in science. Here we investigate the emergence, the rates of emergence and convergence, and the Coding-theorem like behaviour of AP in Turing-subuniversal models of computation. We investigate empirical distributions of computing models in the Chomsky hierarchy. We introduce measures of algorithmic probability and algorithmic complexity based upon resource-bounded computation, in contrast to previously thoroughly investigated distributions produced from the output distribution of Turing machines. This approach allows for numerical approximations to algorithmic (Kolmogorov-Chaitin) complexity-based estimations at each of the levels of a computational hierarchy. We demonstrate that all these estimations are correlated in rank and that they converge both in rank and values as a function of computational power, despite fundamental differences between computational models. In the context of natural processes that operate below the Turing universal level because of finite resources and physical degradation, the investigation of natural biases stemming from algorithmic rules may shed light on the distribution of outcomes. We show that up to 60\% of the simplicity/complexity bias in distributions produced even by the weakest of the computational models can be accounted for by Algorithmic Probability in its approximation to the Universal Distribution.

More than 50 years ago Bongard introduced 100 visual concept learning problems as a testbed for intelligent vision systems. These problems are now known as Bongard problems. Although they are well known in the cognitive science and AI communities only moderate progress has been made towards building systems that can solve a substantial subset of them. In the system presented here, visual features are extracted through image processing and then translated into a symbolic visual vocabulary. We introduce a formal language that allows representing complex visual concepts based on this vocabulary. Using this language and Bayesian inference, complex visual concepts can be induced from the examples that are provided in each Bongard problem. Contrary to other concept learning problems the examples from which concepts are induced are not random in Bongard problems, instead they are carefully chosen to communicate the concept, hence requiring pragmatic reasoning. Taking pragmatic reasoning into account we find good agreement between the concepts with high posterior probability and the solutions formulated by Bongard himself. While this approach is far from solving all Bongard problems, it solves the biggest fraction yet.

Keisuke Sakaguchi Title: "Robust Text Correction for Grammar and Fluency" Abstract: Robustness has always been a desirable property for natural language processing. In many cases, NLP models (e.g., parsing) and downstream applications (e.g., MT) perform poorly when the input contains noise such as spelling errors, grammatical errors, and disfluency. In this talk, I will present three recent results on error correction models: character, word, and sentence level respectively. For character level, I propose semi-character recurrent neural network, which is motivated by a finding in Psycholinguistics, called Cmabrigde Uinervtisy (Cambridge University) effect. For word-level robustness, I propose an error-repair dependency parsing algorithm for ungrammatical texts.

We present a neural model for representing snippets of code as continuous distributed vectors. The main idea is to represent code as a collection of paths in its abstract syntax tree, and aggregate these paths, in a smart and scalable way, into a single fixed-length \emph{code vector}, which can be used to predict semantic properties of the snippet. We demonstrate the effectiveness of our approach by using it to predict a method's name from the vector representation of its body. We evaluate our approach by training a model on a dataset of $14$M methods. We show that code vectors trained on this dataset can predict method names from files that were completely unobserved during training. Furthermore, we show that our model learns useful method name vectors that capture semantic similarities, combinations, and analogies. Comparing previous techniques over the same data set, our approach obtains a relative improvement of over $75\%$, being the first to successfully predict method names based on a large, cross-project, corpus.