Current Projects

Social Robotics Workshop

Dates: 2010 – 2013

Funding: NCWIT

This workshop presents basic computer science concepts by focusing both on the core fundamentals of robotics and the role of social interaction in future robot platforms. This workshop engages students with interesting tasks that explore the capabilities and ideas behind robotic platforms. We try to meet students “where they are”, allowing them to explore robotic concepts in terms of exercises based around every day real life scenarios. Prior tasks explored in early versions of this workshop included home aides to help with brushing their hair, and robots designed to assist in buying ice cream and fetching it from the freezer.

Past Projects

CPAM: Changing Positions, Altering Minds

Dates: 2010 – 2011

Funding: Lockheed Martin

The objective of this project is to develop computational models that can assist those engaged in influence operations across social, cultural, and diacultural boundaries. The main hypothesis is the notion that social and cultural groups and communities can be defined by networks of concepts (ethical, social, religious, political, etc.) that are recognized by the members of these communities along with the valuation system associated with these concepts. Within those networks, there are concepts which are ‘vulnerable’ to change, if presented with sufficient new or conflicting information with respect to the valuations already in the network.

DSAMRD: Detecting Social Actions and Roles in Multiparty Dialogue


Dates: 2009 – 2012

Funding: iARPA

The DSARMD project (Detecting Social Actions and Roles in Multiparty Dialogue) investigates the dynamics of small group interactions across media, cultures and languages. Specifically, we are developing computational models of how certain social phenomena such as leadership, power, and conflict are signaled and reflected in language through the choice of lexical, syntactic, semantic and conversational forms by discourse participants in face-to-face discourse and in on-line interaction. Our objective is to build a prototype system that given a representative segment of multiparty task-oriented dialogue would automatically detect, with a high degree of accuracy, those social and cultural elements in language and discourse structure that are commonly associated with (a) conflict and disagreement, (b) power and dominance and (c) leadership.

ECO: Effective Communication Online

Dates: 2009 – 2010

Funding: Lockheed Martin

The purpose of this research was to develop methods that could assess the effectiveness of communications destined for members of specific diacultural groups, as exemplified by their collective presence on Internet social media, specifically blogs. We collected Internet blog material representing opposing cultural groups, and containing opposing value systems. Analysis of the collected data led to construction of a simple, stochastic model for representing these value systems. Subsequently, improvements in the models were made using machine learning approaches, which enable us to capture community specific valuation models, and facilitate comparison between new messages and the target community, reporting mismatches in valuations of concepts between the two.

DeER: Deliberative E-Rule Making Decision Facilitation Project


Dates: 2007 – 2010

Funding: National Science Foundation

By exploiting advances in question answering technology, discourse tracking, topic detection and summarization, we will present a suite of tools to enable online, collaborative deliberative decision making, providing a facilitation agent able to overcome traditional limitations in the online deliberative process.

Social Robotics


Dates: 2007 – 2010

Funding: National Science Foundation

This awards brings together researchers and educators from UAlbany’s ILS, Rensselaer Polytechnic Institute, Schenectady County Community College, Union College and the Schenectady Museum and Suits-Bueche Planetarium. Students will utilize robots as a platform to learn key computer science concepts, begin to program, and learn about both hardware and software and the interplay between the two as part of a multi-school effort to deliver a unique educational experience for the Capital Region.



Dates: 2006 – 2011

Funding: EU 6th Framework Program

To make computer interfaces more human-like has been a longstanding goal of human-computer interaction research, which has influenced all major areas of Multimodal Interfaces, from speech understanding to emotion recognition. This endeavour has given rise to the concept of agent-based interfaces, ones in which the user communicates with a virtual character using multimodal dialogue.

These are often referred to as Embodied Conversational Agents (ECAs), to emphasise how the use of a character could focus natural language interaction as well as the affective component of the interface. The context of the COMPANIONS project concerns ECAs as persistent companions for users in carrying out activities within their personal digital space, such as organising image and text records of their lives as coherent narratives. This will require extending multimodal dialogue capabilities to support the social and affective aspects of interaction, as well as to give them specific cognitive capabilities, so as to support their relation to the user and organise information. The emotional aspects of the interface will be similarly grounded in the linguistic performance of human-computer dialogue, extended to such aspects as politeness and humour.

COLLANE: The Collaborative Analytical Environment


Dates: 2006 – 2008

Funding: DTO

This project aims at developing advanced information access and analysis tools that exploit the strength of collaborative analysis through the interactive support environment where individual users can take advantage not only of the system’s capabilities to rapidly filter and locate relevant information, but also from each other’s actions and insights. The goal is to create an environment where groups of analysts can work together effectively on complex intelligence problems.

HITIQA Phase III: High-Quality Interactive Question Answering


Dates: 2006 – 2008

Funding: DTO

HITIQA is a question-answering system, driven by natural language human-computer dialogue. HITIQA is sponsored by the Advanced Research Development Activity (ARDA), which is an intelligence community center for conducting advanced research and development related to information technology. HITIQA is also a component of ARDA’s Advanced Question and Answering for Intelligence (AQUAINT). Presently, intelligence analysts could use current query systems that search databases through a document retrieval system with structured queries, but none of these systems would equal the convenience and directness of HITIQA. HITIQA was developed by a team of researchers at the University at Albany, and Rutgers University.

AMITIES: Automated Multilingual Interaction with Information and Services


Dates: 2001 – 2005

Funding: EU 5th Framework Program & DARPA

This international Consortium consisting of EU- and US-based partners, has developed novel technologies for building empirically induced dialogue processors to support multilingual human-computer interaction and integrate these technologies into systems for call centre applications.

The central aspect of this work has been the synthesis of dialogue systems from large corpora of human-human and human-computer dialogues. The project extended techniques that have been successful for creating robust parsers and speech recognisers to the area of dialogue systems by utilising probabilistic models and learning technologies. Our claim is that such methodologies have matured to the point that they can be applied to the efficient and reliable creation of robust dialogue systems. Once such technologies are operative at the dialogue level, they will be applicable to multiple languages, to varieties of applications, and to different user groups.



Dates: 2002

Funding: Department of Work and Pensions & Xansa

Essentially a consultancy with DWP to bring natural language technology to the worktrain website. Using a language-based search box, we are trying to empower users of the site who are looking for work and training opportunities, by giving them new, natural methods to interact with the data sources.

NAMIC: News Agencies Multilingual Information Categorisation


Dates: 2000 – 2003

Funding: EU 5th Framework Program

NAMIC exploits state-of-art human language technologies for content-driven news classification and user-driven enrichment, based on the intelligent extraction of information relevant to the user. The project’s goal is to harness advanced Natural Language Processing (NLP) technologies for multilingual news customisation and broadcasting and to develop tools to near-market products.

YPA: Yellow Pages Assistant


Dates: 1998 – 2000

Funding: British Telecommunications Research Lab

The YPA is a natural language question answering system with dialogue component, aimed at giving users precise answers to queries about Yellow Pages advertisers.

SNAP: Simple Natural Language Access Platform


Dates: 1996 – 1998

Funding: BICC

Part of the SALT initiative, we were working with an industrial partner, BICC to create a natural language front end to query both a relational database (Oracle), and a text base (OpenText Livelink Intranet).