This podcast follows from the presentation at https://youtu.be/LxFUB3uvosw. In this episode, we're joined by Alexander Visheratin, AI researcher at Playground AI. In this talk, Alexander Visheratin explores NLLB-CLIP: the current state of multilingual multimodal models, the challenges faced in their development, and his experience creating one of the top CLIP-like models that supports 200 languages.

In celebration of World Tuberculosis (TB) Day, our esteemed guest, Zainab Abdullahi, a seasoned Nigerian medical lab scientist with extensive hands-on experience with TB patients, delves into the topic of tuberculosis in Africa. Join us as we examine the challenges posed by TB infection, its prevalence in Africa, and explore its causes, prevention, and treatment options.🎥:  https://youtu.be/zEqBPhVRHio“The World TB day is commemorated to raise public awareness about the devastating health, social and economic consequences of tuberculosis (TB) and to step up efforts to end the global TB epidemic.” – https://www.who.int/campaigns/world-tb-day/Learn more about Zainab AbdullahiLinkedln: https://www.linkedin.com/in/zainab-abdullahi-16a241203--- Send in a voice message: https://podcasters.spotify.com/pod/show/lanfrica-talks-podcast/message

There is a lot of misunderstanding and hype surrounding AI, fueled by exaggerated marketing and capitalist motives. This episode features Kathryn Kure and Jonathan Poritz as we delve into 'AInt': their critical examination of AI. Together, we explore the harms and exaggerations surrounding AI, including its unsustainable energy consumption, exclusivity, and ethical dilemmas, while also assessing its potential threat to the global creative commons.Watch their original presentation at the Creative Commons Global Summit 2023: https://www.youtube.com/watch?v=8fPQz5ZCKqQ&list=PLPDjhzarwzChlePJkWR1iOaoUpCDQYGXt&index=21Learn more about Jonathan PortitzMastodon: @poritzj@mastodon.social  Website: https://poritz.net/jonathanLearn more about Kathryn KureTwitter: @datamynaLinkedln: https://www.linkedin.com/in/kathrynkure/--- Send in a voice message: https://podcasters.spotify.com/pod/show/lanfrica-talks-podcast/message

This podcast follows from the presentation at https://youtu.be/UN1pk2lvME8. In this episode, we're joined by Chisom Paula Ogamba, sociolinguist, discourse analyst, feminist and lecturer. In this episode, we chat with Chisom Ogamba about Nigerian language and cultural habits and how they influence sexual abuse and victim silence in Nigeria.--- Send in a voice message: https://podcasters.spotify.com/pod/show/lanfrica-talks-podcast/message

This podcast follows from the presentation at https://youtu.be/B6ARayTDmz4. In this episode, we're joined by Brian Muhia, co-Founder of Fahamu Inc, and CTO of Equiano Institute’s Causality, Interpretability and Representation Learning group. We talk about the struggles of charting your own path (in a world where paths to success are usually predefined). We round it up with discussions around mechanistic interpretability, causality and deep learning.--- Send in a voice message: https://podcasters.spotify.com/pod/show/lanfrica-talks-podcast/message

This podcast follows from the presentation at https://youtu.be/rV5bUikOe88. In this episode, we're joined by Opeyemi Osakuade, a PhD student at the University of Edinburgh, engaged in the study of speech models, with a particular focus on low-resourced languages. Opeyemi Lukman discusses the challenges of applying advanced speech representation models, like HuBERT, to tonal languages, using Yorùbá as a case study. The episode explores the strengths and limitations of HuBERT in capturing the nuances of Yorùbá's tonal system.Follow Opeyemi OsakuadeTwitter: https://twitter.com/o_feranmi_Linkedln: https://www.linkedin.com/in/opeyemi-osakuade/--- Send in a voice message: https://podcasters.spotify.com/pod/show/lanfrica-talks-podcast/message

This podcast follows from the presentation at https://youtu.be/J9CNrw4OevQ. In this episode, we're joined by Lukman Enegi Ismaila, a machine learning engineer and neuroimaging researcher. Lukman discusses his innovative African Biobank solution, tackling the scarcity of medical imagery data for African scientific research. Lukman envisions a collaborative Biobank framework uniting organizations, governments, health centers, academia, and technology institutions. Tune in for a glimpse into the future of transformative health research in Africa!--- Send in a voice message: https://podcasters.spotify.com/pod/show/lanfrica-talks-podcast/message

This podcast follows from the presentation at https://youtu.be/F0Hrhu9-E-M. In this episode, we host Elizabeth Salesky, PhD student at Johns Hopkins University and ☕😀 caffeinated optimist. We discuss building AI evaluation datasets under realistic conditions.--- Send in a voice message: https://podcasters.spotify.com/pod/show/lanfrica-talks-podcast/message

This podcast follows from the presentation at https://youtu.be/1wqbhOIX6I4. In this episode, we host Favour James and discuss topics like healthcare patient data digitization, data accessibility and sharing, cancer research at the Google Summer of Code, building datasets, and being an “artificial intelligence engineer” in the future.--- Send in a voice message: https://podcasters.spotify.com/pod/show/lanfrica-talks-podcast/message

This conversation follows from the presentation at https://youtu.be/7RYG9X_VE-Y. In this podcast episode, we engage in a comprehensive discussion with Denise DiPersio (the Associate Director at the Linguistic Data Consortium (LDC)), touching upon various subjects, including the history of LDC, its membership structure, the distinctive licensing procedure it employs, the Data Scholarship program, and the how and why of sharing African language resources through LDC.--- Send in a voice message: https://podcasters.spotify.com/pod/show/lanfrica-talks-podcast/message

This episode follows from the presentation at https://youtu.be/M0X53u9qBpk. In this podcast episode, we delve into a discussion with Graham Morrissey, COO at Way With Words, covering a range of topics: digital equity and inclusivity, the role of data governance, ethics and community,  fostering data collaborations, towards a sustainable community of passionate data workers (incentivization, maintaining the drive), opportunities, linking passion and profit oriented burgeoning communities, challenges and considerations for data creation and digital language technologies in Africa.--- Send in a voice message: https://podcasters.spotify.com/pod/show/lanfrica-talks-podcast/message

This episode follows from the presentation at https://youtu.be/JoaZoy4fRPs. In this episode, we delve into a discussion with Professor Yoshua Bengio covering a range of themes like intelligence and setting goals, superhuman AI, AI (mis)alignment, need to ensure AI democracy, governmental steps to be taken, open-source AI, and Bayesian agents. 🎧🎥 Subscribe for more conversations from Lanfrica Talks.--- Send in a voice message: https://podcasters.spotify.com/pod/show/lanfrica-talks-podcast/message

This episode follows from the presentation at https://youtu.be/z1ubta3RBp8. In this episode we chat with Paul Okewunmi where we navigate the synergy of research and deployment in Jumia Lens, explore the DSN OAU community, delve into ML mentorship, the Deep Learning Indaba's global collaboration, and the transformative value of sharing your work.--- Send in a voice message: https://podcasters.spotify.com/pod/show/lanfrica-talks-podcast/message

In the latest episode (Ep. 21) of Lanfrica Talks, we had the honor of hosting the seasoned Prof. Vukosi Marivate! 🤩 We delved into some captivating discussions about African language technologies, data sharing, and data governance. Vukosi shared valuable insights on empowering communities in the realm of Natural Language Processing (NLP). Moreover, we explored the significance of grounded research with real-life impact, highlighting how technology can truly uplift and transform societies. One of the highlights was chatting about the remarkable Vuk'uzenzele South African Multilingual Corpus and its role in advancing language technologies in Africa. If you're curious about NLP or looking for inspiration to embark on your own journey in this field, this episode is a must-watch! 🎧📺 Stay tuned for more thought-provoking conversations on Lanfrica Talks.--- Send in a voice message: https://podcasters.spotify.com/pod/show/lanfrica-talks-podcast/message

We were pleased to host Aidan Pine and Patrick Littell for episode 20 of our Lanfrica Talks. They motivated and presented ReadAlong Studio, an application for performing language zero-shot (L0-shot) text-speech alignment. The ReadAlong Studio Web Interface is designed to create digital interactive "readalong" audio books that highlight words as they are spoken and allow users to replay individual words when clicked. The ReadAlong Studio Web Interface is open-source, user-friendly, prioritizes privacy and data sovereignty, allows for a variety of standard export formats, and is designed to work for the majority of the world's written and spoken languages. Listen and learn interesting concepts like “pay-up-front problem”, “language-level zero-shot applications”, data sovereignty and respecting user’s data while building language technologies. Watch on YouTube: https://youtu.be/MrRq3JD5D40--- Send in a voice message: https://podcasters.spotify.com/pod/show/lanfrica-talks-podcast/message

In this episode, Shester Gueuwou discusses his award-winning research on machine translation for African sign languages. He describes his journey into this research, including the hype surrounding sign language technologies and its negative impact, as well as the distinction between recognition and translation in sign languages. Sign language translation is an active area of research with the main goal of bridging the communication gap between deaf and hearing individuals. But while there has been significant progress for sign languages from high-income countries, minimal research has been conducted on African sign language translation. --- Send in a voice message: https://podcasters.spotify.com/pod/show/lanfrica-talks-podcast/message

In this episode, we host babs.ai, a personalized assistant service that prioritizes productivity and provides support for African users who communicate in low-resource languages. (think ChatGPT but for Africa). Temi Babs, the company's founder, introduces the platform, its unique technologies and services, and their most recent advancements in virtual assistants and chatbots.--- Send in a voice message: https://podcasters.spotify.com/pod/show/lanfrica-talks-podcast/message

In this episode, Batazia's co-founder and CTO, Ndipabonga Atanga, discussed the development of speech recognition technology for African languages. She explored the exciting developments in speech recognition technology for African languages on mobile applications at Batazia and discussed, among other things, how speech recognition technology is enhancing accessibility to various services such as education, healthcare, and e-commerce.--- Send in a voice message: https://podcasters.spotify.com/pod/show/lanfrica-talks-podcast/message

In this episode, Ife Adebara takes us through the AfroLID toolkit project, a neural LID toolkit for 517 African languages and varieties. Language identification (LID) is a crucial precursor for NLP, especially for mining web data. Problematically, most of the world’s 7000+ languages today are not covered by LID technologies. AfroLID exploits a multi-domain web dataset manually curated from across 14 language families utilizing five orthographic systems. When evaluated on our blind Test set, AfroLID achieves 95.89 F_1-score. We also compare AfroLID to five existing LID tools that each cover a small number of African languages, finding it to outperform them on most languages. We further show the utility of AfroLID in the wild by testing it on the acutely under-served Twitter domain. Finally, we offer a number of controlled case studies and perform a linguistically-motivated error analysis that allow us to both showcase AfroLID’s powerful capabilities and limitations.--- Send in a voice message: https://podcasters.spotify.com/pod/show/lanfrica-talks-podcast/message

Language Documentation (LD) has gained popularity in African linguistics in recent years, driven by efforts to conserve endangered languages and make language data more accessible. However, the impact of LD on African linguistics research is questionable, as data-driven research is still limited in this field. In this special podcast episode, Dr Emmanuel Ngue Um discusses the role of LD as a new heuristics, not just for data collection but for studying the intersection of social life, places, and objects in language reality. The use of digital technologies and Artificial Intelligence can help to re-examine the representation of linguistic diversity in Africa and its potential impact on traditional African linguistics.--- Send in a voice message: https://podcasters.spotify.com/pod/show/lanfrica-talks-podcast/message