Hey Siri, Can You Hear Me? #ASA184

Hey Siri, Can You Hear Me? #ASA184

Experiments show how speech and comprehension change when people communicate with artificial intelligence.

Media Contact:
Ashley Piccone
AIP Media
301-209-3090
media@aip.org

CHICAGO, May 9, 2023 – Millions of people now regularly communicate with AI-based devices, such as smartphones, speakers, and cars. Studying these interactions can improve AI’s ability to understand human speech and determine how talking with technology impacts language.

In their talk, “Clear speech in the new digital era: Speaking and listening clearly to voice-AI systems,” Georgia Zellou and Michelle Cohn of the University of California, Davis will describe experiments to investigate how speech and comprehension change when humans communicate with AI. The presentation will take place Tuesday, May 9, at 12:40 p.m. Eastern U.S. in the Los Angeles/Miami/Scottsdale room, as part of the 184th Meeting of the Acoustical Society of America running May 8-12 at the Chicago Marriott Downtown Magnificent Mile Hotel.

Humans change their voice when communicating with AI. Credit: Michelle Cohn

In their first line of questioning, Zellou and Cohn examined how people adjust their voice when communicating with an AI system compared to talking with another human. They found the participants produced louder and slower speech with less pitch variation when they spoke to voice-AI (e.g., Siri, Alexa), even across identical interactions.

On the listening side, the researchers showed that how humanlike a device sounds impacts how well listeners will understand it. If a listener thinks the voice talking is a device, they are less able to accurately understand. However, if it sounds more humanlike, their comprehension increases. Clear speech, like in the style of a newscaster, was better understood overall, even if it was machine-generated.

“We do see some differences in patterns across human- and machine-directed speech: People are louder and slower when talking to technology. These adjustments are similar to the changes speakers make when talking in background noise, such as in a crowded restaurant,” said Zellou. “People also have expectations that the systems will misunderstand them and that they won’t be able to understand the output.”

Clarifying what makes a speaker intelligible will be useful for voice technology. For example, these results suggest that text-to-speech voices should adopt a “clear” style in noisy conditions.

Looking forward, the team aims to apply these studies to people from different age groups and social and language backgrounds. They also want to investigate how people learn language from devices and how linguistic behavior adapts as technology changes.

“There are so many open questions,” said Cohn. “For example, could voice-AI be a source of language change among some speakers? As technology advances, such as with large language models like ChatGPT, the boundary between human and machine is changing – how will our language change with it?”

———————– MORE MEETING INFORMATION ———————–
Main meeting website: https://acousticalsociety.org/asa-meetings/
Technical program: https://eppro02.ativ.me/web/planner.php?id=ASASPRING23&proof=true

ASA PRESS ROOM
In the coming weeks, ASA’s Press Room will be updated with newsworthy stories and the press conference schedule at https://acoustics.org/asa-press-room/.

LAY LANGUAGE PAPERS
ASA will also share dozens of lay language papers about topics covered at the conference. Lay language papers are 300 to 500 word summaries of presentations written by scientists for a general audience. They will be accompanied by photos, audio, and video. Learn more at https://acoustics.org/lay-language-papers/.

PRESS REGISTRATION
ASA will grant free registration to credentialed and professional freelance journalists. If you are a reporter and would like to attend the meeting or virtual press conferences, contact AIP Media Services at media@aip.org.  For urgent requests, AIP staff can also help with setting up interviews and obtaining images, sound clips, or background information.

ABOUT THE ACOUSTICAL SOCIETY OF AMERICA
The Acoustical Society of America (ASA) is the premier international scientific society in acoustics devoted to the science and technology of sound. Its 7,000 members worldwide represent a broad spectrum of the study of acoustics. ASA publications include The Journal of the Acoustical Society of America (the world’s leading journal on acoustics), JASA Express Letters, Proceedings of Meetings on Acoustics, Acoustics Today magazine, books, and standards on acoustics. The society also holds two major scientific meetings each year. See https://acousticalsociety.org/.

Artificial intelligence in music production: controversy & opportunity

Joshua Reiss Reiss – joshua.reiss@qmul.ac.uk
Twitter: @IntelSoundEng

Queen Mary University of London, Mile End Road, London, England, E1 4NS, United Kingdom

Popular version of 3aSP1-Artificial intelligence in music production: controversy and opportunity, presented at the 183rd ASA Meeting.

Music production
In music production, one typically has many sources. They each need to be heard simultaneously, but can all be created in different ways, in different environments and with different attributes. The mix should have all sources sound distinct yet contribute to a nice clean blend of the sounds. To achieve this is labour intensive and requires a professional engineer. Modern production systems help, but they’re incredibly complex and all require manual manipulation. As technology has grown, it has become more functional but not simpler for the user.

Intelligent music production
Intelligent systems could analyse all the incoming signals and determine how they should be modified and combined. This has the potential to revolutionise music production, in effect putting a robot sound engineer inside every recording device, mixing console or audio workstation. Could this be achieved? This question gets to the heart of what is art and what is science, what is the role of the music producer and why we prefer one mix over another.

Artificial Intelligence Figure 1: The architecture of an automatic mixing system. [Image courtesy of the author] Figure 1 Caption: The architecture of an automatic mixing system. [Image courtesy of the author]

Perception of mixing
But there is little understanding of how we perceive audio mixes. Almost all studies have been restricted to lab conditions; like measuring the perceived level of a tone in the presence of background noise. This tells us very little about real world cases. It doesn’t say how well one can hear lead vocals when there are guitar, bass and drums.

Best practices
And we don’t know why one production will sound dull while another makes you laugh and cry, even though both are on the same piece of music, performed by competent sound engineers. So we needed to establish what is good production, how to translate it into rules and exploit it within algorithms. We needed to step back and explore more fundamental questions, filling gaps in our understanding of production and perception.

Knowledge engineering
We used an approach that incorporated one of the earliest machine learning methods, knowledge engineering. Its so old school that its gone out of fashion. It assumes experts have already figured things out, they are experts after all. So let’s capture best practices as a set of rules and processes. But this is no easy task. Most sound engineers don’t know what they did. Ask a famous producer what he or she did on a hit song and you often get an answer like ‘I turned the knob up to 11 to make it sound phat.” How do you turn that into a mathematical equation? Or worse, they say it was magic and can’t be put into words.

We systematically tested all the assumptions about best practices and supplemented them with listening tests that helped us understand how people perceive complex sound mixtures. We also curated multitrack audio, with detailed information about how it was recorded, multiple mixes and evaluations of those mixes.

This enabled us to develop intelligent systems that automate much of the music production process.

Video Caption: An automatic mixing system based on a technology we developed.

Transformational impact
I gave a talk about this once in a room that had panel windows all around. These talks are usually half full. But this time it was packed, and I could see faces outside pressed up against the windows. They all wanted to find out about this idea of automatic mixing. It’s  a unique opportunity for academic research to have transformational impact on an entire industry. It addresses the fact that music production technologies are often not fit for purpose. Intelligent systems open up new opportunities. Amateur musicians can create high quality mixes of their content, small venues can put on live events without needing a professional engineer, time and preparation for soundchecks could be drastically reduced, and large venues and broadcasters could significantly cut manpower costs.

Taking away creativity
Its controversial. We entered an automatic mix in a student recording competition as a sort of Turing Test. Technically we cheated, because the mixes were supposed to be made by students, not by an ‘artificial intelligence’ (AI) created by a student. Afterwards I asked the judges what they thought of the mix. The first two were surprised and curious when I told them how it was done. The third judge offered useful comments when he thought it was a student mix. But when I told him that it was an ‘automatic mix’, he suddenly switched and said it was rubbish and he could tell all along.

Mixing is a creative process where stylistic decisions are made. Is this taking away creativity, is it taking away jobs? Such questions come up time and time again with new technologies, going back to 19th century protests by the Luddites, textile workers who feared that time spent on their skills and craft would be wasted as machines could replace their role in industry.

Not about replacing sound engineers
These are valid concerns, but its important to see other perspectives. A tremendous amount of music production work is technical, and audio quality would be improved by addressing these problems. As the graffiti artist Banksy said “All artists are willing to suffer for their work. But why are so few prepared to learn to draw?”

Creativity still requires technical skills. To achieve something wonderful when mixing music, you first have to achieve something pretty good and address issues with masking, microphone placement, level balancing and so on.

Video Caption: Time offset (comb filtering) correction, a technical problem in music production solved by an intelligent system.

The real benefit is not replacing sound engineers. Its dealing with all those situations when a talented engineer is not available; the band practicing in the garage, the small restaurant venue that does not provide any support, or game audio, where dozens of sounds need to be mixed and there is no miniature sound engineer living inside the games console.

1aBA5 – AI and the future of pneumonia diagnosis

Xinliang Zheng – lzheng@intven.com
Sourabh Kulhare – skulhare@intven.com
Courosh Mehanian — cmehanian@intven.com
Ben Wilson — bwilson@intven.com
Intellectual Ventures Laboratory
14360 SE Eastgate Way
Bellevue, WA 98007, U.S.A.

Zhijie Chen – chenzhijie@mindray.com
SHENZHEN MINDRAY BIO-MEDICAL ELECTRONICS CO., LTD.
Mindray Building, Keji 12th Road South,High-tech Industrial Park,
Nanshan, Shenzhen 518057, P.R. China

Popular version of paper 1aBA5
Presented Monday morning, November 5, 2018
176th ASA Meeting, Minneapolis, MN

A key gap for underserved communities around the world is the lack of clinical laboratories and specialists to analyze samples. But thanks to advances in machine learning, a new generation of ‘smart’ point-of-care diagnostics are filling this gap and, in some cases, even surpassing the effectiveness of specialists at a lower cost.

Take the case of pneumonia. Left untreated, pneumonia can be fatal. The leading cause of death among children under the age of five, pneumonia claims the lives of approximately 2,500 a day – nearly all of them in low-income nations.

To understand why, consider the differences in how the disease is diagnosed in different parts of the world. When a doctor in the U.S. suspects a patient has pneumonia, the patient is usually referred to a highly-trained radiologist, who takes a chest X-ray using an expensive machine to confirm the diagnosis.

Because X-ray machines and radiologists are in short supply across much of sub-Saharan Africa and Asia and the tests themselves are expensive, X-ray diagnosis is simply not an option for the bottom billion. In those settings, if a child shows pneumonia symptoms, a cough and a fever, she is usually treated with antibiotics as a precautionary measure and sent on her way. If, in fact, the child does not have pneumonia, this means she receives unnecessary antibiotics, leaving her untreated for her real illness and putting her health at risk. The widespread overuse of antibiotics also contributes to the buildup in resistance of the so-called “superbug” – a global threat.

In this context, an interdisciplinary team of algorithm developers, software engineers and global health experts at Intellectual Ventures’ Global Good—a Bill and Melinda Gates-backed technology fund that invents for humanitarian impact—considered the possibility of developing a low-cost tool capable of automating pneumonia diagnosis.

The team turned to ultrasound – an affordable, safe, and widely-available technology that can be used to diagnose pneumonia with a comparable level of accuracy to X-ray.

It wouldn’t be easy. To succeed, the device would need to be cost-effective, portable, easy-to-use and able to do the job quickly, accurately and automatically in challenging environments.

Global Good started by building an algorithm to recognize four key features associated with lung conditions in an ultrasound image – pleural line, B-line, consolidation and pleural effusion. This called for convolutional neural networks (CNNs)—a machine learning method well-suited for image classification tasks. The team trained the algorithm by showing it ultrasound images collected from over 70 pediatric and adult patients. The features were annotated on the images by expert sonographers to ensure accuracy.

Figure 1: Pleural line (upper arrow) and a-lines (lower arrow), indication of normal lung

pneumonia

Figure 2: Consolidation (upper arrow) and merged B-line (lower arrow), indication of abnormal lung fluid and potentially pneumonia

Early tests show that the algorithm can successfully recognize abnormal lung features in ultrasound images and those features can be used to diagnose pneumonia as reliably as X-ray imaging—a highly encouraging outcome.

The algorithm will eventually be installed on an ultrasound device and used by minimally-trained healthcare workers to make high-quality diagnosis accessible to children worldwide at the point of care. Global Good hopes that the device will eventually bring benefits to patients in wealthy markets as well, in the form of a lower-cost, higher quality and faster alternative to X-ray.