Two of the world's leading AI speech researchers unite at aiOla to fuel and spearhead innovation efforts
TEL AVIV, Israel, Oct. 23, 2024 /PRNewswire/ -- aiOla, a global leader in advanced AI speech technology, announced today that Prof. Dr. Bhiksha Raj, a renowned automatic speech recognition researcher with over 550 published papers, is joining aiOla as their Distinguished Scientist. Dr. Raj will work closely with aiOla's Chief Scientist, Prof. Dr. Yossi Keshet, one of the top-cited spoken language processing researchers globally. The collaboration of these two longtime friends and experts at aiOla marks a significant milestone, setting the stage for major advancements in AI-powered speech innovation.
Academia has long been the bedrock of AI research, fostering foundational theories and driving long-term innovations. However, breakthroughs over the last decade, such as DeepMind's AlphaGo (2015) and the Transformer architecture (2017), led to advancements we see today, like ChatGPT. These advancements have shifted the AI landscape, with industry now outpacing academia in cutting-edge research and development. This trend is also reflected in the movement of talent between sectors. In 2019, 13% of new AI faculty positions in North America were filled by professionals moving from industry to academia; by 2022, this figure had dropped to 7%. While academia continues to contribute valuable research, industry is increasingly taking the lead in driving practical AI advancements and applications.
Dr. Bhiksha Raj, a pioneer in speech recognition, has over 30 years of experience in the field. His career includes nearly a decade at Mitsubishi Research Labs, followed by over fifteen years as a professor at Carnegie Mellon University (CMU). During this time, Dr. Raj played a pivotal role in developing CMU Sphinx, which was one of the earliest and most widely used open-source speech recognition systems. With over 2.5 million downloads, CMU Sphinx has been instrumental in advancing speech technology and has served as the foundation for numerous companies in the field.
"I've been in the field since the 1990s and have witnessed the evolution of automatic speech recognition," said Dr. Bhiksha Raj, Distinguished Scientist at aiOla. "We're entering a new era of AI built on the foundation laid by academia, with industries now accelerating AI's advancement and delivering its benefits directly to the public. It's exciting times and I couldn't be more thrilled to be a part of aiOla. aiOla is at the forefront of this transformation, consistently delivering impactful research and innovative products that are setting new standards in industries from manufacturing to aviation and beyond."
Dr. Yossi Keshet, with nearly 25 years of experience in the field, is an award-winning automatic speech recognition researcher and professor at the Technion. Dr. Keshet has made profound contributions to the automatic speech recognition community, authoring over 100 academic papers. At aiOla, Dr. Keshet has led several projects that have set new industry benchmarks, such as creating AdaKWS, an AI keyword spotting model that outperforms OpenAI's Whisper in accurately transcribing company-specific jargon, and KG-Whisper, a state-of-the-art mechanism to combine keyword spotting with a transformer-based speech recognizer.
"Speech recognition is one of the most exciting fields to be in during this AI renaissance," said Dr. Yossi Keshet, Chief Scientist at aiOla. "At aiOla, my team and I are continuing to break barriers, providing real-world impact to our enterprise clients. I am excited to welcome Bhiksha to aiOla. As one of the field's most published AI speech researchers, Bhiksha's extensive knowledge and background are invaluable, and our friendship will serve as a constant source of motivation and inspiration. Together, we will push the limits of what AI speech technology can do and take aiOla to new heights."
Dr. Raj first met Dr. Keshet in 2009 at a conference where they discussed Dr. Keshet's work on keyword spotting. The two researchers formed an immediate friendship and have since collaborated on several projects, including research on steganography and adversarial robustness. While the two will continue their professorships at CMU and at the Technion, they will collaborate at aiOla to drive innovation and create cutting-edge speech recognition solutions.
"I am particularly eager to work alongside not only a respected colleague but also a dear friend, Yossi Keshet," said Dr. Bhiksha Raj, Distinguished Scientist at aiOla. "I've known of Yossi's work since 2005 when I was developing word spotting and wake-up word detection for my Sphinx system. His new algorithm on max-margin structured models immediately transformed my system. His contributions have been transformative, and today, most companies utilizing keyword spotting or wake-word detection – the technology behind voice commands like 'Hey Siri' or 'Alexa' – are likely employing some variant of his research."
"Welcoming Bhiksha to aiOla marks a significant chapter in our story," said Alon Peleg, Chief Operating Officer of aiOla. "Bhiksha is joining a team of the brightest minds in AI speech technology. Yossi's and Bhiksha's collaboration will not only drive innovation at aiOla forward by advancing our intellectual property and patent portfolio but also push the entire AI speech research community to new heights. We're thrilled to have another world-class expert join our ranks, further solidifying aiOla's position at the forefront of speech technology innovation."
About aiOla:
aiOla's patented technology comprehends over 120 languages, and discerns jargon, abbreviations and acronyms, demonstrating a low error rate even in noisy environments. aiOla's technology converts manual processes in critical industries into data-driven, paperless, AI-powered workflows through cutting-edge speech recognition.
Contact
Ali Goldberg
Concrete Media for aiOla
[email protected]
Image - https://mma.prnewswire.com/media/2538202/Dr_Bhiksha_Raj.jpg
Image - https://mma.prnewswire.com/media/2538203/Dr_Yossi_Keshet.jpg
SOURCE aiOla
Share this article