A team of researchers from the University of Trento (Trento, Italy) has been working with IBM to advance the state of the art in teaching computers to conversationally interact with humans. The Trento group is led by Professor Giuseppe Riccardi and Professor Alessandro Moschitti. Their work furthers the ability of computers to understand and use natural human speech.
Riccardi says the team is “investigating how to interpret natural language queries in terms that the machine can work with to find candidate answers out of many possible answers. The important aspect of this,” says Riccardi, “is that we are teaching machines how to do this, rather than storing patterns or deterministic rules.”
The gulf between the specificity of medical terminology and the generality of lay medical speech will pose no special problems, asserts Riccardi. “This discontinuity in language is the same that we observe in conversational and written language,” he says. “The speech and language research community has been working on this problem for many years already.”
Going forward, Professor Riccardi says, “One of the major challenges is to compound Watson’s cognitive ability to understand massive amounts of documents into communicative and social skills. Ultimately this will have tremendous impact on sectors such as personalised healthcare. The health domain has (to date) been marginally touched by the latest advances in speech and language technologies,” he continues. “We expect future speech and language technology will be more pervasive in human doctor and patient interactions, and in company processes” such as Electronic Health Record (EHR) management.
If winning at chess is viewed as the first milestone in machine intelligence and the Jeopardy! Challenge is seen as the second, the next major milestone, Professor Riccardi says, is to “give social and interactive skills to artificial agents.” In other words, for a machine to converse with a human as if it were a human.
Prof. Dr.-Ing. Giuseppe Riccardi is head of the Signals and Interactive Systems Lab of the Department of Information Engineering and Computer Science (DISI) at the University of Trento, and is affiliated with the interdisciplinary Department of Information and Communication Technology and the Center for Mind/Brain Sciences. He is the founder and director of the Adaptive Multimodal Information and Interfaces (AMI2) Lab.