IBM Research-IISc Workshop on Knowledge and Learning

IBM Research and IISc are organizing a workshop on Knowledge and Learning, to celebrate 20 years of IBM Research in India. The event features talks by researchers from IISc and IBM Research on frontier areas in Artificial Intelligence and Machine Learning. The topics include Knowledge Extraction and Representation, Deep Learning, Summarization, Automated Question Answering among others.

~~Apply here to attend the workshop. Deadline: 5pm IST, Mar 2, 2018 (Fri). Notification: Mar 3, 2018.~~

Schedule

	Wednesday, 7 March 2018
09:15 - 9:30	Introduction & Welcome
09:30 - 10:00	IBM Research - IISc Collaboration Overview : Vinayaka D Pandit
10:00 - 11:00	Keynote Talk: Pushpak Bhattacharya (IIT Patna/Bombay) Advances in Sentiment Analysis [SLIDES]
	Abstract: This talk will present our multi-faceted and long standing work on sentiment analysis, in the general setting of the currently highly important area of text analytics. Huge amount of electronic data is in textual form and lot of it is in the form of opinion, stances, emotion epithets. We take 3 advanced topics of sentiment analysis- Sarcasm, Thwarting and Eye-Tracking- that make in-roads into computing that straddles semantics and pragmatics. We try to capture and use INCONGRUITY for sarcasm processing. A sentence like "I love being ignored" is sarcastic, because it has incongruity in it- "love" is a positive sentiment word, while "ignore" is negative. We exploit incongruity and many other features- traditional and novel- to detect sarcasm automatically. Thwarting too is a challenging problem, as the presence of a critical negative aspect thwarts good effects of many positive aspects. Ontology is at the heart of such processing. Finally, A new line of investigation by us is the use of eye tracking features for sentiment detection: "Eyes give away what words do not tell". Many of our systems based on rich set of features and techniques including SVM, Deep Learning etc. report accuracies better than existing values. The talk is based on work done with graduate students (Balamurali, Raksha, Aditya, Abhijeet, Kevin and others) that has been reported frequently in ACL, EMNLP, AAAI and such fora.
11:00 - 11:30	Coffee and Tea
11:30 - 12:00	Talk 1: Ashish R Mittal (IBM IRL) Natural Language Querying over Relational Data Stores
	Abstract: We present ATHENA, an ontology-driven system for natural language querying of complex relational databases. Natural language interfaces to databases enable users easy access to data, without the need to learn a complex query language, such as SQL. ATHENA uses domain specific ontologies, which describe the semantic entities, and their relationships in a domain. We propose a unique two-stage approach, where the input natural language query (NLQ) is first translated into an intermediate query language over the ontology, called OQL, and subsequently translated into SQL. Our two-stage approach allows us to decouple the physical layout of the data in the relational store from the semantics of the query, providing physical independence. Moreover, ontologies provide richer semantic information, such as inheritance and membership relations, that are lost in a relational schema. By reasoning over the ontologies, our NLQ engine is able to accurately capture the user intent. (Joint work with Diptikalyan Saha, Avrilia Floratou, Karthik Sankaranarayanan, Umar Farooq Minhas, Fatma Özcan)
12:00 - 12:30	Talk 2: Venkatesh Babu (CDS, IISc) Adversarial Attacks on Deep Models
	Abstract: Deep learning systems are shown to be vulnerable to adversarial noise: small but structured perturbation added to the input that affects the model’s prediction drastically. The most successful Deep Neural Network based object classifiers have also been observed to be susceptible to adversarial attacks with almost imperceptible perturbations. More importantly, the adversarial perturbations exhibit cross model generalizability. That is, the perturbations learned on one model can fool another model even if the second model has a different architecture or has been trained on a disjoint subset of training images. These adversarial noise poses a severe threat for deploying ML based systems in the real world. Particularly, for the applications that involve safety and privacy (e.g., autonomous driving and access granting). In this talk we will discuss some of the recent attempts to create such perturbations, especially universal adversarial perturbations (UAPs).
12:30 - 14:00	Lunch (CSA Lawns)
14:00 - 14:30	Talk 3: Chiranjib Bhattacharya (CSA, IISc) OWL to the rescue of LASSO [SLIDES]
	Abstract: Penalizing a model by L1 norm, also known as LASSO penalty, has proven to be a powerful tool for learning sparse models in several settings including linear regression. However, in linear regression LASSO fails to recover the true model if the predictors are correlated. This is an important open problem and has sparked new interest in investigating alternatives to LASSO which can provably guarantee the discovery of true model. In this talk we will discuss Ordered weighted L1 norm(OWL) and show how it can discover the true model even in the presence of strong correlation.
14:30 - 15:00	Talk 4: Amrita Saha (IBM IRL) Complex Sequential Question Answering
	Abstract: While conversing with chatbots, humans typically tend to ask many questions, a significant portion of which can be answered by referring to large-scale knowledge graphs (KG). While Question Answering (QA) and dialog systems have been studied independently, there is a need to study them closely to evaluate such real-world scenarios faced by bots involving both these tasks. Towards this end, we introduce the task of Complex Sequential QA which combines the two tasks of (i) answering factual questions through complex inferencing over a realistic-sized KG of millions of entities, and (ii) learning to converse through a series of coherently linked QA pairs. Through a labor intensive semi-automatic process, involving in-house and crowdsourced workers, we created a dataset containing around 200K dialogs with a total of 1.6M turns. Further, unlike existing large scale QA datasets which contain simple questions that can be answered from a single tuple, the questions in our dialogs require a larger subgraph of the KG. Specifically, our dataset has questions which require logical, quantitative, and comparative reasoning as well as their combinations. This calls for models which can: (i) parse complex natural language questions, (ii) use conversation context to resolve coreferences and ellipsis in utterances, (iii) ask for clarifications for ambiguous queries, and finally (iv) retrieve relevant subgraphs of the KG to answer such questions. However, our experiments with a combination of state of the art dialog and QA models show that they clearly do not achieve the above objectives and are inadequate for dealing with such complex real world settings. We believe that this new dataset coupled with the limitations of existing models as reported in this paper should encourage further research in Complex Sequential QA.
15:00 - 15:30	Coffee and Tea
15:30 - 16:00	Talk 5: Anirban Laha (IBM IRL) Generating Natural Language Descriptions from Structured Data
	Abstract: We present our recent work focussing on the task of generating natural language descriptions from a structured table of facts containing fields (such as nationality, occupation, etc) and values (such as Indian, actor, director, etc). One simple choice is to treat the table as a sequence of fields and values and then use a standard seq2seq model for this task. However, such a model is too generic and does not exploit task specific characteristics. For example, while generating descriptions from a table, a human would attend to information at two levels: (i) the fields (macro level) and (ii) the values within the field (micro level). Further, a human would continue attending to a field for a few timesteps till all the information from that field has been rendered and then never return back to this field (because there is nothing left to say about it). To capture this behavior we used (i) a fused bifocal attention mechanism which exploits and combines this micro and macro level information and (ii) a gated orthogonalization mechanism which tries to ensure that a field is remembered for a few timesteps and then forgotten. We experimented with a recently released dataset which contains fact tables about people and their corresponding one line biographical descriptions in English. Our experiments showed that the proposed model gives 21% relative improvement over a recently proposed state of the art method and 10 % relative improvement over basic seq2seq models.
16:00 - 16:30	Talk 6: Partha Talukdar (CDS, IISc) Canonicalizing Open Knowledge Graphs
	Abstract: Open Information Extraction (OpenIE) methods extract (noun phrase, relation phrase, noun phrase) triples from text, resulting in the construction of large Open Knowledge Bases (Open KBs). The noun phrases (NPs) and relation phrases in such Open KBs are not canonicalized, leading to the storage of redundant and ambiguous facts. Recent research has posed canonicalization of Open KBs as clustering over manually-defined feature spaces. Manual feature engineering is expensive and often sub-optimal. In order to overcome this challenge, we propose Canonicalization using Embeddings and Side Information (CESI) – a novel approach which performs canonicalization over learned embeddings of Open KBs. CESI extends recent advances in KB embedding by incorporating relevant NP and relation phrase side information in a principled manner. Through extensive experiments on multiple real-world datasets, we demonstrate CESI’s effectiveness.
16:30 - 17:00	Closure

Contacts: Sreyash Kenkre (srekenkr [at] in.ibm.com), Partha Talukdar (ppt [at] iisc.ac.in)

IBM Research-IISc Workshop on Knowledge and Learning

Celebrating 20 years of IBM Research India

Schedule