The face of computing is constantly changing. Wearable computers in the form of glasses or watches are becoming more and more common. These devices have very small screens (measured in millimeters), and users often interact with them through voice input and audio feedback. Weather is one of the most regularly checked app category on smart devices, but weather results on these devices are often limited to raw data, canned responses, or sentence templates with numbers plugged in. The goal for this project was to build a system that could generate weather forecast text, which could then be read to a user through text-to-speech. By using methods in language generation, the system can generate weather forecast text in millions of different ways. This is all computed locally, and it covers every possible weather case. In order to generate natural weather forecast texts, the system retrieved raw weather data from a weather API and created the text through six methods: content determination, document structuring, sentence aggregation, lexical choice, referring expression generation, and text realization. Content determination is the process of deciding on what information to include in a computer generated text. The document structuring phase deals with the order and structure of the information. Sentence aggregation is the merging of similar sentences to improve readability and to reduce redundancy. Lexical choice is the process of putting words to concepts. Referring expression generation is the process of identifying objects, regions, time periods, and locations within a text. Finally text realization involves creating sentences with proper syntax, morphology, and orthography. Through these six stages, a system was developed that could generate unique weather forecast text from raw data accurately and efficiently. It was built for iOS devices with Apple's new programming language, Swift, and it will be ported to the Apple Watch when the API is fully opened to developers.

The objective of this research is to determine an approach for automating the learning of the initial lexicon used in translating natural language sentences to their formal knowledge representations based on lambda-calculus expressions. Using a universal knowledge representation and its associated parser, this research attempts to use word alignment techniques to align natural language sentences to the linearized parses of their associated knowledge representations in order to learn the meanings of individual words. The work includes proposing and analyzing an approach that can be used to learn some of the initial lexicon.

LPMLN is a recent probabilistic logic programming language which combines both Answer Set Programming (ASP) and Markov Logic. It is a proper extension of Answer Set programs which allows for reasoning about uncertainty using weighted rules under the stable model semantics with a weight scheme that is adopted from Markov Logic. LPMLN has been shown to be related to several formalisms from the knowledge representation (KR) side such as ASP and P-Log, and the statistical relational learning (SRL) side such as Markov Logic Networks (MLN), Problog and Pearl’s causal models (PCM). Formalisms like ASP, P-Log, Problog, MLN, PCM have all been shown to embeddable in LPMLN which demonstrates the expressivity of the language. Interestingly, LPMLN has also been shown to reducible to ASP and MLN which is not only theoretically interesting, but also practically important from a computational point of view in that the reductions yield ways to compute LPMLN programs utilizing ASP and MLN solvers. Additionally, the reductions also allow the users to compute other formalisms which can be reduced to LPMLN.

This thesis realizes two implementations of LPMLN based on the reductions from LPMLN to ASP and LPMLN to MLN. This thesis first presents an implementation of LPMLN called LPMLN2ASP that uses standard ASP solvers for computing MAP inference using weak constraints, and marginal and conditional probabilities using stable models enumeration. Next, in this thesis, another implementation of LPMLN called LPMLN2MLN is presented that uses MLN solvers which apply completion to compute the tight fragment of LPMLN programs for MAP inference, marginal and conditional probabilities. The computation using ASP solvers yields exact inference as opposed to approximate inference using MLN solvers. Using these implementations, the usefulness of LPMLN for computing other formalisms is demonstrated by reducing them to LPMLN. The thesis also shows how the implementations are better than the native solvers of some of these formalisms on certain domains. The implementations make use of the current state of the art solving technologies in ASP and MLN, and therefore they benefit from any theoretical and practical advances in these technologies, thereby also benefiting the computation of other formalisms that can be reduced to LPMLN. Furthermore, the implementation also allows for certain SRL formalisms to be computed by ASP solvers, and certain KR formalisms to be computed by MLN solvers.

In this thesis, I propose a new technique of Aligning English sentence words

with its Semantic Representation using Inductive Logic Programming(ILP). My

work focusses on Abstract Meaning Representation(AMR). AMR is a semantic

formalism to English natural language. It encodes meaning of a sentence in a rooted

graph. This representation has gained attention for its simplicity and expressive power.

An AMR Aligner aligns words in a sentence to nodes(concepts) in its AMR

graph. As AMR annotation has no explicit alignment with words in English sentence,

automatic alignment becomes a requirement for training AMR parsers. The aligner in

this work comprises of two components. First, rules are learnt using ILP that invoke

AMR concepts from sentence-AMR graph pairs in the training data. Second, the

learnt rules are then used to align English sentences with AMR graphs. The technique

is evaluated on publicly available test dataset and the results are comparable with

state-of-the-art aligner.

In recent years, several methods have been proposed to encode sentences into fixed length continuous vectors called sentence representation or sentence embedding. With the recent advancements in various deep learning methods applied in Natural Language Processing (NLP), these representations play a crucial role in tasks such as named entity recognition, question answering and sentence classification.

Traditionally, sentence vector representations are learnt from its constituent word representations, also known as word embeddings. Various methods to learn the distributed representation (embedding) of words have been proposed using the notion of Distributional Semantics, i.e. “meaning of a word is characterized by the company it keeps”. However, principle of compositionality states that meaning of a sentence is a function of the meanings of words and also the way they are syntactically combined. In various recent methods for sentence representation, the syntactic information like dependency or relation between words have been largely ignored.

In this work, I have explored the effectiveness of sentence representations that are composed of the representation of both, its constituent words and the relations between the words in a sentence. The word and relation embeddings are learned based on their context. These general-purpose embeddings can also be used as off-the- shelf semantic and syntactic features for various NLP tasks. Similarity Evaluation tasks was performed on two datasets showing the usefulness of the learned word embeddings. Experiments were conducted on three different sentence classification tasks showing that our sentence representations outperform the original word-based sentence representations, when used with the state-of-the-art Neural Network architectures.

Answer Set Programming (ASP) is one of the main formalisms in Knowledge Representation (KR) that is being widely applied in a large number of applications. While ASP is effective on Boolean decision problems, it has difficulty in expressing quantitative uncertainty and probability in a natural way.

Logic Programs under the answer set semantics and Markov Logic Network (LPMLN) is a recent extension of answer set programs to overcome the limitation of the deterministic nature of ASP by adopting the log-linear weight scheme of Markov Logic. This thesis investigates the relationships between LPMLN and two other extensions of ASP: weak constraints to express a quantitative preference among answer sets, and P-log to incorporate probabilistic uncertainty. The studied relationships show how different extensions of answer set programs are related to each other, and how they are related to formalisms in Statistical Relational Learning, such as Problog and MLN, which have shown to be closely related to LPMLN. The studied relationships compare the properties of the involved languages and provide ways to compute one language using an implementation of another language.

This thesis first presents a translation of LPMLN into programs with weak constraints. The translation allows for computing the most probable stable models (i.e., MAP estimates) or probability distribution in LPMLN programs using standard ASP solvers so that the well-developed techniques in ASP can be utilized. This result can be extended to other formalisms, such as Markov Logic, ProbLog, and Pearl’s Causal Models, that are shown to be translatable into LPMLN.

This thesis also presents a translation of P-log into LPMLN. The translation tells how probabilistic nonmonotonicity (the ability of the reasoner to change his probabilistic model as a result of new information) of P-log can be represented in LPMLN, which yields a way to compute P-log using standard ASP solvers or MLN solvers.

Background: With the large amount of pharmacological and biological knowledge available in literature, finding novel drug indications for existing drugs using in silico approaches has become increasingly feasible. Typical literature-based approaches generate new hypotheses in the form of protein-protein interactions networks by means of linking concepts based on their cooccurrences within abstracts. However, this kind of approaches tends to generate too many hypotheses, and identifying new drug indications from large networks can be a time-consuming process.

Methodology: In this work, we developed a method that acquires the necessary facts from literature and knowledge bases, and identifies new drug indications through automated reasoning. This is achieved by encoding the molecular effects caused by drug-target interactions and links to various diseases and drug mechanism as domain knowledge in AnsProlog, a declarative language that is useful for automated reasoning, including reasoning with incomplete information. Unlike other literature-based approaches, our approach is more fine-grained, especially in identifying indirect relationships for drug indications.

Conclusion/Significance: To evaluate the capability of our approach in inferring novel drug indications, we applied our method to 943 drugs from DrugBank and asked if any of these drugs have potential anti-cancer activities based on information on their targets and molecular interaction types alone. A total of 507 drugs were found to have the potential to be used for cancer treatments. Among the potential anti-cancer drugs, 67 out of 81 drugs (a recall of 82.7%) are indeed known cancer drugs. In addition, 144 out of 289 drugs (a recall of 49.8%) are non-cancer drugs that are currently tested in clinical trials for cancer treatments. These results suggest that our method is able to infer drug indications (original or alternative) based on their molecular targets and interactions alone and has the potential to discover novel drug indications for existing drugs.

Due to vast resources brought by social media services, social data mining has

received increasing attention in recent years. The availability of sheer amounts of

user-generated data presents data scientists both opportunities and challenges. Opportunities are presented with additional data sources. The abundant link information

in social networks could provide another rich source in deriving implicit information

for social data mining. However, the vast majority of existing studies overwhelmingly

focus on positive links between users while negative links are also prevailing in real-

world social networks such as distrust relations in Epinions and foe links in Slashdot.

Though recent studies show that negative links have some added value over positive

links, it is dicult to directly employ them because of its distinct characteristics from

positive interactions. Another challenge is that label information is rather limited

in social media as the labeling process requires human attention and may be very

expensive. Hence, alternative criteria are needed to guide the learning process for

many tasks such as feature selection and sentiment analysis.

To address above-mentioned issues, I study two novel problems for signed social

networks mining, (1) unsupervised feature selection in signed social networks; and

(2) unsupervised sentiment analysis with signed social networks. To tackle the first problem, I propose a novel unsupervised feature selection framework SignedFS. In

particular, I model positive and negative links simultaneously for user preference

learning, and then embed the user preference learning into feature selection. To study the second problem, I incorporate explicit sentiment signals in textual terms and

implicit sentiment signals from signed social networks into a coherent model Signed-

Senti. Empirical experiments on real-world datasets corroborate the effectiveness of

these two frameworks on the tasks of feature selection and sentiment analysis.

For the past three decades, the design of an effective strategy for generating poetry that matches that of a human’s creative capabilities and complexities has been an elusive goal in artificial intelligence (AI) and natural language generation (NLG) research, and among linguistic creativity researchers in particular. This thesis presents a novel approach to fixed verse poetry generation using neural word embeddings. During the course of generation, a two layered poetry classifier is developed. The first layer uses a lexicon based method to classify poems into types based on form and structure, and the second layer uses a supervised classification method to classify poems into subtypes based on content with an accuracy of 92%. The system then uses a two-layer neural network to generate poetry based on word similarities and word movements in a 50-dimensional vector space.

The verses generated by the system are evaluated using rhyme, rhythm, syllable counts and stress patterns. These computational features of language are considered for generating haikus, limericks and iambic pentameter verses. The generated poems are evaluated using a Turing test on both experts and non-experts. The user study finds that only 38% computer generated poems were correctly identified by nonexperts while 65% of the computer generated poems were correctly identified by experts. Although the system does not pass the Turing test, the results from the Turing test suggest an improvement of over 17% when compared to previous methods which use Turing tests to evaluate poetry generators.

There have been extensive research in how news and twitter feeds can affect the outcome of a given stock. However, a majority of this research has studied the short term effects of sentiment with a given stock price. Within this research, I studied the long-term effects of a given stock price using fundamental analysis techniques. Within this research, I collected both sentiment data and fundamental data for Apple Inc., Microsoft Corp., and Peabody Energy Corp. Using a neural network algorithm, I found that sentiment does have an effect on the annual growth of these companies but the fundamentals are more relevant when determining overall growth. The stocks which show more consistent growth hold more importance on the previous year’s stock price but companies which have less consistency in their growth showed more reliance on the revenue growth and sentiment on the overall company and CEO. I discuss how I collected my research data and used a multi-layered perceptron to predict a threshold growth of a given stock. The threshold used for this particular research was 10%. I then showed the prediction of this threshold using my perceptron and afterwards, perform an f anova test on my choice of features. The results showed the fundamentals being the better predictor of stock information but fundamentals came in a close second in several cases, proving sentiment does hold an effect over long term growth.