The communication between user and computer is more and more shifting towards voice interfaces. Voice interfaces are the most natural form of communication to most users. They can be used in a variety of situations (e.g., while driving), even from a distance. The Cornell Database Group is working on leveraging voice query interfaces for manipulating and analyzing structured data.

Our research focuses on three challenges. First, resolving ambiguities in voice input, due to noisy speech recognition. Second, extracting high-level trends from query results, suitable for concise output. Third, reducing computational overheads by exploiting particularities of voice interfaces (e.g., by overlapping computation with voice output). Our most recent paper (talk video below) uses pre-processing to summarize query results with minimal overheads.

Publications

• VLDB’21 Robust voice querying with MUVE: optimally visualizing results of phonetically similar queries. Ziyun Wei, Immanuel Trummer, and Connor Anderson.
• SIGMOD’21 Demonstrating robust voice querying with MUVE: optimally visualizing results of phonetically similar queries. Ziyun Wei, Immanuel Trummer, Connor Anderson.
• ICDE’21 Optimally summarizing data by small fact sets for concise answers to voice queries. Immanuel Trummer, Connor Anderson.
• VLDB’20 Demonstrating the voice-based exploration of large data sets with CiceroDB-Zero. Immanuel Trummer.
• CIDR’19 Data vocalization with CiceroDB. Immanuel Trummer.
• SIGMOD’19 A holistic approach for query evaluation and result vocalization in voice-based OLAP. Immanuel Trummer, Yicheng Wang, Saketh Mahankali.
• VLDB’18 Vocalizing large time series efficiently. Immanuel Trummer, Mark Bryan, and Ramya Narasimha.
• VLDB’17 Optimizing voice-based output of relational data. Immanuel Trummer, Jiancheng Zhu, and Mark Bryan.

Awards

• CIDA Grant for “Co-creating a human-machine interface better adapted for on-farm data recording, curation, management, and use”.
• Mark Bryan wins honorable mention for the CRA Outstanding Undergraduate Researcher Award.
• Google Faculty Research Award for project “optimizing voice-based output of relational data”.

Funding

We thankfully acknowledge funding by the following organizations:

• Cornell Institute for Digital Agriculture
• J. P. Morgan
• Google
• Lockheed Martin