In Pt. 2 of my series on TikTok, I continue to delve deeper into TikTok's algorithms to explore the question of whether they introduce and propagate bias. I'll also share the results of a survey I conducted to assess whether people could detect bias in the news they consume on social media. Finally, I'll share my opinion on policy recommendations to reduce bias in social media platforms. 

To explore technology bias, I recently did a research project about TikTok, a popular social media platform. In this Part 1 of a two part series, I'll introduce the background and history of TikTok as well as some key important industry terms like shadowbanning and filter bubbles. I'll also talk about how TikTok's algorithms work to set up a deeper exploration into bias in Part 2. 

I continue my conversation with Dr, Sue Hertz, digging deeper into the complexities of text-to-speech technology, including the challenges of embedding emotions into speech. We talk about specific examples of speech nuance - like the difference in the sound of the letter “p” in the word “Poke” versus “Spoke”. We also discuss how it has taken years of research to get to the point where Synfonica’s text-to-speech system is today.  I loved hearing her talk about her fascination with how listeners parse the nuances of speech. 

Come join me as I speak with Dr. Sue Hertz, Adjunct Professor at Cornell University and President of Synfonica LLC, a speech technology company developing knowledge-based text-to-speech systems and related applications. She shares what initially sparked her interest in linguistics and how learning computer programming became an essential part of her journey. We’ll learn about her pioneering research and development in text-to-speech synthesis over the last several decades. I was especially inspired to hear about how one of her systems is still one of the most widely used programs in screen readers for blind individuals.  

Dr. Hertz asked me to clarify that when she said that the knowledge-based approach Synfonica is taking to text-to-speech synthesis differs from the approaches taken by others, she meant “others today.” Decades ago, there was no choice but to use such an approach due to the severe memory limitations of computer systems at that time. 

I continue my conversation with Professor Cisneros as we delve further into his interest in minority languages such as Mixtec and Guaymi. We discuss the many reasons languages can become extinct, including globalization and colonialism. We also learn about the only language that has ever been truly revitalized! (Can you guess what it is?) 

Come join me as I speak with Professor Carlos Cisneros, former Linguistics professor at Berkeley.  In this podcast, we talk about how his interest in Mesoamerican culture sparked his interest in the study of linguistics. We also discuss  the importance of studying minority languages, and what they teach us about what the human mind is capable of, and how language can reflect upon both universal as well as divergent aspects of our cultures. 

"The fire destroyed the apartment complex."
"People curled up next to the fire with their cocoa."
The word "fire" here means something very different within the context of each of the broader stories, in a way that a dictionary cannot capture. Join me as I continue my conversation with Professor Erk. We discuss the concept of 'story salads' and how they can help us understand how language works, and what it means for a word to have meaning.  We also talk about some fascinating real world implications, such as how we can better ensure accuracy in the social media processing of natural disasters in real time. 

Come join me as I speak with Professor Katrin Erk, a professor of Computational Linguistics at the University of Texas at Austin. We explore what sparked her initial interest in this field, as well as her specialty in computational semantics. We learn about the challenges of understanding the meaning of sentences, and how formal logical structures and machine learning can be very helpful in certain situations, but can't always handle more nuanced or longer stretches of text. I was especially intrigued by how Professor Erk illustrates these challenges through her research on how advice in various on-line forums can take on different and nuanced forms! 

Come join me as I continue my discussion with Professor Elsner. In this podcast we learn about the connection between language and vision, and how language can help us understand visual systems and vice versa. What does the popular children's game Where's Waldo tell us about how we describe and process visual cues? What is the concept of visual salience and how does it relate to language? We discuss the Gricean process and how we make inferences from what is unspoken, and how this concept extends to how we describe what we see. 

Come join me as I speak with Professor Micha Elsner, a professor of Computational Linguistics at Ohio State University. We explore his intellectual journey and what sparked his interest in this field. We also dig into one of his areas of focus - child language acquisition. What is known about how children learn language and what is still a mystery? Are certain language easier to acquire? What is a language 'subsystem'? And how does the 'labor of  lifetimes' continually help us understand more about how languages are learned?