HRL Speech Seminar - Talk of the Town
Speech technology experts from around Israel and across the globe gathered for a full day seminar dedicated to advanced research into speech technologies.
Haifa, Israel - Last week, the leading speech technologies experts from around Israel gathered to fill the auditorium at the IBM Haifa Research Lab (HRL). HRL researchers hosted the leadership seminar, which gathered members of the speech community for a full day of lectures, demos, poster presentations, and the all-important networking. People who attended the seminar remarked on the high quality of talks and were reluctant to leave at the end of the day-continuing to hold heated discussions long after the official closing remarks. David Nahamoo, IBM speech expert and a central figure in the global speech technology domain, delivered the keynote address, discussing the evolution of speech recognition over the years and the new vision for the so-called superhuman speech recognition.
"One of the goals of the seminar was to highlight the leading-edge speech work being done in IBM Research-especially in the area of speech recognition, text-to-speech, and speech analytics," explained Ron Hoory, Manager of Speech Technologies at the Haifa Research Lab and one of the seminar organizers. "Most of the topics focused on speech as a source of knowledge and information that can be analyzed to bring us even more value."
Despite the fact that Hebrew is not a very popular language on a worldwide scale, a significant number of Israeli companies are involved in speech technologies. This is probably due to Israel's expertise in telephony and pioneering work in voice over IP.
In Israel, IBM is expanding its activities in the area of speech technologies and stands out as a definitive industry leader. Aside from the leading-edge research being done at the Haifa Research Lab, a team of developers at the IBM Israel Software Lab (ILSL) has begun several projects that will integrate speech technologies into IBM's unified communications technologies. A Global Business Services team in IBM Israel focuses on providing solutions for the many call centers set up in Israel.
Ron Hoory's team in Haifa is focusing primarily on text-to-speech technologies, with much of their work being used for embedded text-to-speech solutions such as those in today's cars. As road safety rises on our agenda, these systems are becoming more critical since voice is the safest and most convenient interface drivers can use to interact with their cars. Getting GPS coordinates, driving instructions, information on nearby restaurants, or traffic reports is just the tip of the iceberg.
The team's second area of focus revolves around speech analytics, where speech from various sources-whether recorded speech, video clips, or phone calls-is analyzed. This information can then be used, for example, to identify the speaker, pinpoint the speaker's emotions, or search for spoken terms. For example, speech analytics can be used to process conversations and identify who said what. Another use for speech analytics is for call centers; customer speech can be monitored to determine when they are getting frustrated or upset. Researchers at the Haifa Lab are currently involved in two European projects that employ speech analytics: SAPIR (Search on Audio-visual content using Peer-to-peer Information Retrieval) and HERMES (Cognitive Care and Guidance for Active Aging).
"People want the future to arrive already," noted Hoory. "Although we already have very advanced automated systems, we still have lots to do before HAL and Star Trek become a reality."