PeopleofSuperintelligence

Researcher

Giuseppe Riccardi

Giuseppe Riccardi is a computer scientist at the University of Trento whose research covers language modeling, spoken language understanding and human-machine dialogue. Trained in electrical engineering at the University of Padua, where he earned a PhD in 1995, he spent 1993 to 2005 at AT&T Bell Laboratories and AT&T Labs-Research, working on the "How May I Help You?" dialogue program and on active learning for speech recognition. In 2005, he moved to Trento, where he founded the Signals and Interactive Systems Lab and led the university's team that contributed to IBM's Watson. IEEE elected him a Fellow in 2010, and ISCA did so in 2017. He is a full professor and directs the lab, whose recent work tests how well large language models and vision-language models reason.

Timeline

  1. 1990–1995

    Degrees from Padua and Milan

    From 1990 to 1993, he collaborates with Alcatel-Telettra Research Laboratories in Italy. In 1991, he earns a Laurea in electrical engineering from the University of Padua and a master's in information technology from CEFRIEL/Polytechnic of Milan. He receives his PhD in electrical engineering from Padua's Department of Electrical Engineering in 1995.

  2. 1993–1996

    Researcher at AT&T Bell Laboratories

    He joins AT&T Bell Laboratories in Murray Hill, New Jersey, as a researcher in 1993. With colleagues, he designs the AT&T spoken language system that ranks first in the 1994 DARPA ATIS evaluation, and he receives a Bell Labs President Award in 1994.

  3. 1996–2005

    Distinguished Researcher at AT&T Labs

    From 1996 to 2005, he is a Distinguished Researcher in the Speech and Language Processing Lab at AT&T Labs-Research in New Jersey. He and his colleagues carry out spontaneous-speech research for the "How May I Help You?" program, and a 2002 paper he coauthors reports a 38% cut in call-type classification errors for AT&T's HMIHY natural dialog system using word confusion networks. His work on finite-state automata and transducers also leads to Anuvaad, the first large-scale finite-state chain decoder for machine translation.

  4. 2003–2009

    Active learning for speech recognition

    In 2003, he and Dilek Hakkani-Tür describe a method that combines active and unsupervised learning for automatic speech recognition, reducing the labeled data needed for a given word accuracy by 75%. He receives the 2009 IEEE Signal Processing Society Best Paper Award for the article "Active Learning: Theory and Applications to Automatic Speech Recognition".

  5. 2005–2010

    Associate professor at Trento

    He joins the University of Trento's Department of Information Engineering and Computer Science as an associate professor in 2005, and there he founds and directs the Signals and Interactive Systems Lab. The European Commission names him a Marie Curie Excellence Leader for 2006–2010. He leads Trento's team that contributes to IBM's Watson, the machine that wins the Jeopardy! challenge, and serves as General Chair of the IEEE ASRU conference in 2009.

  6. 2010–2023

    IEEE, ISCA and AAIA fellowships

    IEEE elects him a Fellow in 2010 for contributions to algorithms for automatic speech recognition and spoken language processing. He receives an Amazon Alexa Award in 2016, and ISCA names him a Fellow in 2017 for contributions to language modeling, spoken language understanding and dialogue systems. In 2023, he becomes a Fellow of the AAIA.

  7. 2018–2022

    Dialogue systems for mental health

    He becomes a full professor of computer science and information engineering at Trento in 2018. From 2018 to 2022, his research group leads the design and training of human-machine dialogue systems for mental health in the Horizon 2020 project COADAPT. The group builds TEO (Therapy Empowerment Opportunity), a personalized conversational agent evaluated in a registered randomized controlled trial.

  8. As of October 2026

    Head of Trento's SISL lab

    He is a full professor at Trento and directs the Signals and Interactive Systems Lab. In 2025, his group audits the reasoning benchmarks SocialIQa, FauxPas-EAI and ToMi and publishes CIVET, a framework for evaluating vision-language models with controlled stimuli, in Findings of EMNLP 2025. A 2025 study of counting in vision-language models finds that fine-tuning only the output layer improves accuracy by up to 21%.

  1. Google had the most people shaping AI every year from 2014 to 2025Our estimates of which organizations employed the most people shaping AI from 1961 to 2025, with a year-by-year chart.
  2. The United States employs an estimated 53% of people in our data with a known employment countryOur estimates of where AI careers are concentrated now and how countries’ shares of dated positions changed over time.
  3. 46% of living people in our data with known birth and employment countries work outside their birth countryOur estimate of cross-border careers in AI, from country of birth and the country of each current organization.
  4. More people shaping AI studied at Massachusetts Institute of Technology than at any other institutionOur estimates of where the people shaping AI studied and where their careers led, from education records in cited biographies.

Colleagues

People named in this biography, and people with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Giuseppe Riccardi yet. Send us a free-licensed photo.

  1. Giuseppe Riccardi | CV | University of Trento webapps.unitn.it. Checked 19 October 2026.
  2. Lecturers | RTTH Summer School on Speech Technology 2017 rtthss2017.blogs.upv.es. Checked 19 October 2026.
  3. ISCA Archive - Improving spoken language understanding using word confusion networks isca-archive.org. Checked 19 October 2026.
  4. ISCA Archive - Active and unsupervised learning for automatic speech recognition isca-archive.org. Checked 19 October 2026.
  5. SISL - Signal and Interactive Systems Lab: COADAPT sisl.disi.unitn.it. Checked 19 October 2026.
  6. SISL - Signal and Interactive Systems Lab sisl.disi.unitn.it. Checked 19 October 2026.
  7. Garbage In, Reasoning Out? Why Benchmark Scores are Unreliable and What to Do About It arxiv.org. Checked 19 October 2026.
  8. [De|Re]constructing VLMs' Reasoning in Counting arxiv.org. Checked 19 October 2026.
  9. CIVET: Systematic Evaluation of Understanding in VLMs - ACL Anthology aclanthology.org. Checked 19 October 2026.