CS&RIT Presents: Research Idea Ring Talks
RIR: A Computational Approach to Sign Language Understanding
When: Thu, November 19, 12:30pm – 1:00pm
Speaker: Tejaswini Ananthanarayana
Abstract: Sign language is the primary mode of communication in the Deaf and Hard-of-Hearing (DHH) communities. Unfortunately, sign language is not as well understood among the non-signing hearing population leading to limited access and services to the DHH community, and also acts as a barrier between non-signing and DHH people. In my Ph.D. thesis, I am working on improving the sequence modeling for sign language translation and understanding by considering different types of sequence models, various input features, and by understanding the semantic relation between the words and the signs. Currently, my research focuses on a popular publicly available German Sign Language dataset.
Bio: Tej is a fifth-year Ph.D. candidate in the Kate Gleason College of Engineering. She obtained her Master's degree in Electrical Engineering also from KGCoE and her Bachelor's degree in Electronics Engineering from the University of Mumbai, India. She is currently a Development Engineer at ON Semiconductor in Rochester. Originally, I started in research from a computer architecture background with a focus on High-Performance Architecture (HPA) developing a customized instruction set architecture for low power consumption but has since changed to AI where she is now developing models for large scale sign language interpretation.
RIR: Challeges in Enhancing Schema Discovery of JSON Documents
When: Thu, November 19, 1:00pm – 1:15pm
Speaker: Justin Namba
Abstract: Schema discovery is finding the structure of data. It helps users understand the meaning of data and write queries to manipulate it. This is typically easy for relational databases, but complex for non-relational (NoSQL) databases with JavaScript Object Notation (JSON) documents. JSON is a representation of documents that contain objects stored in the form of nested key-value pairs. For relational databases, the schema is predefined because the data they contain is structured, but for NoSQL databases, data is usually unstructured or semi-structured. In a collection of JSON documents, the structure of one document can be completely different from another. Several algorithms were developed to discover schemas from JSON documents, but they provide the physical structure and semantic information that is insufficient for data understanding and analysis. In this paper, we enumerate the major techniques used to extract a schema from JSON documents and present some of the next challenges that need to be addressed within the field of JSON schema discovery to enhance the quality of the discovered schemas. These challenges are (1) distinguishing when keys are data or metadata, (2) detecting frequent patterns, and (3) treating objects whose keys are semantically similar with values that are represented in different structures within nested JSON documents. We suggest methods to separate data from metadata, infer sets of new data types, and match semantically similar fields in different structures, which we call dynamic data representation.
Bio: I am a second-year Ph.D. student in the Computer & Information Sciences working with Prof. Michael Mior. I completed my undergrad at RIT and majored in Computer & Information Technologies.
Event Snapshot
When and Where
Who
Open to the Public
Interpreter Requested?
No