Suresh Manandhar

SemEval-2014 Task 4: Aspect Based Sentiment Analysis

Sentiment analysis is increasingly viewed as a vital task both from an academic and a commercial standpoint. The majority of current approaches, however, attempt to detect the overall polarity of a sentence, paragraph, or text span, irrespective of the entities mentioned (e.g., laptops) and their aspects (e.g., battery, screen). SemEval2014 Task 4 aimed to foster research in the field of aspect-based sentiment analysis, where the goal is to identify the aspects of given target entities and the sentiment expressed for each aspect. The task provided datasets containing manually annotated reviews of restaurants and laptops, as well as a common evaluation procedure. It attracted 163 submissions from 32 teams.

SemEval-2015 Task 12: Aspect Based Sentiment Analysis

SemEval-2015 Task 12, a continuation of SemEval-2014 Task 4, aimed to foster research beyond sentenceor text-level sentiment classification towards Aspect Based Sentiment Analysis. The goal is to identify opinions expressed about specific entities (e.g., laptops) and their aspects (e.g., price). The task provided manually annotated reviews in three domains (restaurants, laptops and hotels), and a common evaluation procedure. It attracted 93 submissions from 16 teams.

Extending a Lexical Ontology by a Combination of Distributional Semantics Signatures

Ontologies are a tool for Knowledge Representation that is now widely used, but the effort employed to build an ontology is high. We describe here a procedure to automatically extend an ontology such as WordNet with domain-specific knowledge. The main advantage of our approach is that it is completely unsupervised, so it can be applied to different languages and domains. Our experiments, in which several domain-specific concepts from a book have been introduced, with no human supervision, into WordNet, have been successful.

Stochastic Constraint Programming: A Scenario-Based Approach

To model combinatorial decision problems involving uncertainty and probability, we introduce scenario based stochastic constraint programming. Stochastic constraint programs contain both decision variables, which we can set, and stochastic variables, which follow a discrete probability distribution. We provide a semantics for stochastic constraint programs based on scenario trees. Using this semantics, we can compile stochastic constraint programs down into conventional (non-stochastic) constraint programs. This allows us to exploit the full power of existing constraint solvers. We have implemented this framework for decision making under uncertainty in stochastic OPL, a language which is based on the OPL constraint modelling language [Van Hentenryck et al., 1999]. To illustrate the potential of this framework, we model a wide range of problems in areas as diverse as portfolio diversification, agricultural planning and production/inventory management.

SemEval-2016 task 5 : aspect based sentiment analysis

This paper describes the SemEval 2016 shared task on Aspect Based Sentiment Analysis (ABSA), a continuation of the respective tasks of 2014 and 2015. In its third year, the task provided 19 training and 20 testing datasets for 8 languages and 7 domains, as well as a common evaluation procedure. From these datasets, 25 were for sentence-level and 14 for text-level ABSA; the latter was introduced for the first time as a subtask in SemEval. The task attracted 245 submissions from 29 teams.

Designing an interactive open-domain question answering system

Interactive question answering (QA), where a dialogue interface enables follow-up and clarification questions, is a recent although long-advocated field of research. We report on the design and implementation of YourQA, our open-domain, interactive QA system. YourQA relies on a Web search engine to obtain answers to both fact-based and complex questions, such as descriptions and definitions. We describe the dialogue moves and management model making YourQA interactive, and discuss the architecture, implementation and evaluation of its chat-based dialogue interface. Our Wizard-of-Oz study and final evaluation results show how the designed architecture can effectively achieve open-domain, interactive QA.

Reviewing and Evaluating Automatic Term Recognition Techniques

Automatic Term Recognition (ATR) is defined as the task of identifying domain specific terms from technical corpora. Termhood-basedapproaches measure the degree that a candidate term refers to a domain specific concept. Unithood-basedapproaches measure the attachment strength of a candidate term constituents. These methods have been evaluated using different, often incompatible evaluation schemes and datasets. This paper provides an overview and a thorough evaluation of state-of-the-art ATRmethods, under a common evaluation framework, i.e. corpora and evaluation method. Our contributions are two-fold: (1) We compare a number of different ATRmethods, showing that termhood-basedmethods achieve in general superior performance. (2) We show that the number of independent occurrences of a candidate term is the most effective source for estimating term nestedness, improving ATRperformance.

Unsupervised Learning of Word Segmentation Rules with Genetic Algorithms and Inductive Logic Programming

This article presents a combination of unsupervised and supervised learning techniques for the generation of word segmentation rules from a raw list of words. First, a language bias for word segmentation is introduced and a simple genetic algorithm is used in the search for a segmentation that corresponds to the best bias value. In the second phase, the words segmented by the genetic algorithm are used as an input for the first order decision list learner CLOG. The result is a set of first order rules which can be used for segmentation of unseen words. When applied on either the training data or unseen data, these rules produce segmentations which are linguistically meaningful, and to a large degree conforming to the annotation provided.

Learning multilingual morphology with Clog

The paper presents the decision list learning system Clog and the results of using it to learn nominal inflections of English, Romanian, Czech, Slovene, and Estonian. The dataset used to induce rules for the synthesis and analysis of the inflectional paradigms of nouns and adjectives of these languages is the Multext-East multilingual tagged corpus. The ILP system FoIDL is also applied to the same dataset, and this paper compares the induction methodology and results of the two systems. The experiment shows that the accuracy of the two systems is comparable when using the same training set. However, while FOIDL is, due to efficiency reasons, severely limited in the size of the training set, CLOG does not suffer from such limitations. With the increase of the training set size possible with CLOG, it significantly outperforms FOIDL and learns highly accurate morphological rules.

Phoneme segmentation of speech

In most approaches to speech recognition, the speech signals are segmented using constant-time segmentation, for example into 25 ms blocks. Constant segmentation risks losing information about the phonemes. Different sounds may be merged into single blocks and individual phonemes lost completely. A more satisfactory approach is to attempt to segment the phoneme boundaries from the speech signals and use these boundaries to define blocks. The discrete wavelet transform (DWT) is interesting in the analysis of speech since it is easy to extract parameters which take into account the properties of the human hearing system. The analysis of the power in different frequency bands offers potential for distinguishing the start and end of phonemes. For many boundaries, there is no discernible drop in overall power, and at some frequencies, the power is broadly constant over the lifetime of the phoneme. However, many phonemes exhibit rapid changes in particular subbands which can be used to detect their start and endpoints. In this paper we apply the DWT to speech signals and analyse the resulting power spectrum and its derivatives to locate candidates for the boundaries of phonemes in continuous speech. We compare the results with hand segmentation and constant segmentation over a number of words. The method proves effective for finding most phoneme boundaries