SS 2017

Semantic Parsing


Semantic Parsing

Summer Semester 2017
Prof. Dr. Alexander Koller
Fri 14-16; C71 Seminar Room U15

Start: Friday, April 21

Semantic parsing is the problem of computing, for a given natural-language expression, a formal semantic representation. Traditionally, this was done using hand-written grammars that simultaneously described the syntactic and semantic structure of an expression. More recently, the problem has enjoyed a lot of attention in computational linguistics in a new disguise. The task is now to learn a mapping between strings and semantic representations from data.

There is a growing literature on semantic parsing, which varies e.g. in the choice of semantic representation formalism, the formal tools with which the string-semantics mapping is represented, and whether a grammar is used at all. Annotated data has become available in various different styles. For instance, one recent focus has been the development of graph-banks, in which English sentences have been annotated manually or automatically with graphs representing the sentence meaning.

Structure of the course. In this seminar, we will achieve two things. First, we will learn about the current state of the art in semantic parsing by reading, presenting, and discussing current research paper. Second, we will work in groups to implement our own semantic parser. Depending on how many students want to participate and what their specific interests are, we can combine these two parts in different ways. We will discuss this in the first meeting.

Because the course contains both traditional “seminar” elements (giving a talk, writing a seminar paper) and “project seminar” elements (implementing and evaluating a system), the course can be taken either as a seminar or as a project seminar.

We will coordinate our work using Piazza. Please subscribe to and be active on the Piazza forum for this course.

Requirements for participation. I assume that participants have at least a fundamental knowledge in theoretical semantics, either from the BSc class “Einführung in die Semantik” or from the MSc class “Foundations of Language Science and Technology”. If all your semantics knowledge is from FLST, you will probably get the most out of this course if you take “Semantic Theory” in parallel. For instance, you should be familiar with simply typed lambda calculus, and have a sense of how one can construct lambda terms for simple natural-language expressions. It may be worth reviewing the semantics chapters of the Jurafsky & Martin book before the course starts.

Requirements for course credit. You will give a talk, contribute to an implemented system, and write a seminar paper. The deadline for the seminar paper will be 15 September 2017. We can talk about details in the first meeting.

Important note. If you are interested in attending this course, it would help me if you could email me beforehand, so I can estimate the number of participants and prepare accordingly.