Learning Effective and Interpretable Semantic Models using Non-Negative Sparse Embedding

In this paper, we introduce an application of matrix factorization to produce corpus-derived, distributional
models of semantics that demonstrate cognitive plausibility. We find that word representations
learned by Non-Negative Sparse Embedding (NNSE), a variant of matrix factorization, are sparse,
effective, and highly interpretable. To the best of our knowledge, this is the first approach which
yields semantic representation of words satisfying these three desirable properties. Though extensive
experimental evaluations on multiple real-world tasks and datasets, we demonstrate the superiority
of semantic models learned by NNSE over other state-of-the-art baselines.

Documents and Dependencies: an Exploration of Vector Space Models for Semantic Composition

In most previous research on distributional semantics, Vector Space Models (VSMs) of words are built either from topical information (e.g., documents in which a word is present), or from syntactic/semantic types of words (e.g., dependency parse links of a word in sentences), but not both. In this paper, we explore the utility of combining these two representations to build VSM for the task of semantic composition of adjective-noun phrases. Through extensive experiments on benchmark datasets, we find that even though a type-based VSM is effective for semantic composition, it is often outperformed by a VSM built using a combination of topic- and type-based statistics. We also introduce a new evaluation task wherein we predict the composed vector representation of a phrase from the brain activity of a human subject reading that phrase. We exploit a large syntactically parsed corpus of 16 billion tokens to build our VSMs, with vectors for both phrases and words, and make them publicly available.

Interpretable semantic vectors from a joint model of brain-and text-based meaning

Vector space models (VSMs) represent word meanings as points in a high dimensional space. VSMs are typically created using a large text corpora, and so represent word semantics as observed in text. We present a new algorithm (JNNSE) that can incorporate a measure of semantics not previously used to create VSMs: brain activation data recorded while people read words. The resulting model takes advantage of the complementary strengths and weaknesses of corpus and brain activation data to give a more complete representation of semantics. Evaluations show that the model 1) matches a behavioral measure of semantics more closely, 2) can be used to predict corpus data for unseen words and 3) has predictive power that generalizes across brain imaging technologies and across subjects. We believe that the model is thus a more faithful representation of mental vocabularies.

Using Brain Data for Sentiment Analysis

We present the results of exploratory experiments using lexical valence extracted from brain using electroencephalography (EEG) for sentiment analysis. We selected 78 English words (36 for training and 42 for testing), presented as stimuli to 3 English native speakers. EEG signals were recorded from the subjects while they performed a mental imaging task for each word stimulus. Wavelet decomposition was employed to extract EEG features from the time-frequency domain. The extracted features were used as inputs to a sparse multinomial logistic regression (SMLR) classifier for valence classification, after univariate ANOVA feature selection. After mapping EEG signals to sentiment valences, we exploited the lexical polarity extracted from brain data for the prediction of the valence of 12 sentences taken from the SemEval-2007 shared task, and compared it against existing lexical resources.

Sarcasm Detection on Czech and English Twitter

This paper presents a machine learning approach to sarcasm detection on Twitter in two languages – English and Czech. Although there has been some research in sarcasm detection in languages other than English (e.g., Dutch, Italian, and Brazilian Portuguese), our work is the first attempt at sarcasm detection in the Czech language. We created a large Czech Twitter corpus consisting of 7,000 manually-labeled tweets and provide it to the community. We evaluate two classifiers with various combinations of features on both the Czech and English datasets. Furthermore, we tackle the issues of rich Czech morphology by examining different preprocessing techniques. Experiments show that our language-independent approach significantly outperforms adapted state-of-the-art methods in English (F-measure 0.947) and also represents a strong baseline for further research in Czech (F-measure 0.582).

A Compositional and Interpretable Semantic Space

Vector Space Models (VSMs) of Semantics are useful tools for exploring the semantics of single words, and the composition of words to make phrasal meaning. While many methods can estimate the meaning (i.e. vector) of a phrase, few do so in an interpretable way. We introduce a new method (CNNSE) that allows word and phrase vectors to adapt to the notion of composition. Our method learns a VSM that is both tailored to support a chosen semantic composition operation, and whose resulting features have an intuitive interpretation. Interpretability allows for the exploration of phrasal semantics, which we leverage to analyze performance on a behavioral task.

Evidence of syntactic working memory usage in MEG data

While reading times are often used to measure working memory load, frequency effects (such as surprisal or n-gram frequencies) also have strong confounding effects on reading times. This work uses a naturalistic audio corpus with magnetoencephalographic (MEG) annotations to measure working memory load during sentence processing. Alpha oscillations in posterior regions of the brain have been found to correlate with working memory load in non-linguistic tasks (Jensen et al., 2002), and the present study extends these findings to working memory load caused by syntactic center embeddings. Moreover, this work finds that frequency effects in naturally-occurring stimuli do not significantly contribute to neural oscillations in any frequency band, which suggests that many modeling claims could be tested on this sort of data even without controlling for frequency effects.

Unsupervised Solution Post Identification from Discussion Forums

Discussion forums have evolved into a dependablesource of knowledge to solvecommon problems. However, only a minorityof the posts in discussion forumsare solution posts. Identifying solutionposts from discussion forums, hence, is animportant research problem. In this paper,we present a technique for unsupervisedsolution post identification leveraginga so far unexplored textual feature, thatof lexical correlations between problemsand solutions. We use translation modelsand language models to exploit lexicalcorrelations and solution post characterrespectively. Our technique is designedto not rely much on structural featuressuch as post metadata since suchfeatures are often not uniformly availableacross forums. Our clustering-based iterativesolution identification approach basedon the EM-formulation performs favorablyin an empirical evaluation, beatingthe only unsupervised solution identificationtechnique from literature by a verylarge margin. We also show that our unsupervisedtechnique is competitive againstmethods that require supervision, outperformingone such technique comfortably.

Linguistics and Information Theory: Analytic advantages

The analytic advantages of central concepts from linguistics and information theory, and the analogies demonstrated between them, for understanding patterns of retrieval from full-text indexes to documents are developed. The interaction between the syntagm and the paradigm in computational operations on written language in indexing, searching, and retrieval is used to account for transformations of the signified or meaning between documents and their representation and between queries and documents retrieved. Characteristics of the message, and messages for selection for written language, are brought to explain the relative frequency of occurrence of words and multiple word sequences in documents. The examples given in the companion article are revisited and a fuller example introduced. The signified of the sequence stood for, the term classically used in the definitions of the sign, as something standing for something else, can itself change rapidly according to its syntagm. A greater than ordinary discourse understanding of patterns in retrieval is obtained.