Hardware Podcasts

Google I/O | AI Fairness and Women in Tech9

EE Times Current · 13 May 2019 · 16m

What this episode covers

Rick Merritt and Dylan McGrath discuss Google I/O, highlighting Sundar Pichai's focus on privacy and on-device AI, alongside new Pixel and Home Display hardware. Junko Yoshida introduces EE Times' special project on AI fairness, exploring how biased training data and 'black box' algorithms can lead to discriminatory outcomes in high-stakes areas like lending. Sally Ward Foxton details how financial institutions use machine learning for credit scoring, noting that while tools like S Finance's game theory models exist to tune for fairness, banks often prioritize profit over equity. Finally, Junko and McGrath reflect on VerveCon, a women-focused tech conference, discussing the unique career challenges and resilience required for women in engineering at major tech firms like Google and Microsoft.

Most relevant to

large enterprises

Topics

Mentioned

Companies
Google, Facebook, S Finance, Microsoft, Oracle, LinkedIn, Intel
Products
Pixel, Home Display
Also
Sundar Pichai, Mark Zuckerberg, EE Times, J. Budzick, VerveCon, Suda

Related episodes from other shows

Other shows on the same ground, closest first.

  • Embedded.fm · 1h 07m

    367: Data of Our Lives

    Dr. Ayanna Howard, Dean of Engineering at Ohio State University and former Chair at Georgia Tech, discusses her book 'Sex, Race, and Robots' and the pervasive issue of bias in AI systems. She explains how bias enters the pipeline through data collection, algorithm design, and testing, citing examples like early camera optics excluding darker skin tones and biased resume screening tools. Howard advocates for diverse development teams, third-party auditing standards similar to the FDA, and consumer pressure to drive ethical AI adoption.

    AI & machine learningSecurity & privacy
    7 chapters

  • Embedded.fm · 1h 02m

    306: What Is in the Magic Box?

    Dr. Loretta Cheeks, an AI consultant and founder of the nonprofit Strong Ties, explains her research on detecting implicit bias in unstructured text using machine learning. She details her methodology, which involves analyzing news articles to identify framing and bias intensity, and discusses her background in industry at companies like Honeywell and General Dynamics before earning her PhD. Cheeks also describes her STEAM outreach programs, which use culturally relevant tools like hip hop and coding to engage underserved youth in STEM fields.

    AI & machine learningCareers & education
    7 chapters

  • Embedded.fm · 1h 13m

    252: A Good Heuristic for Pain Tolerance

    Katie Malone, a data scientist at Civis Analytics and host of the Linear Digressions podcast, discusses the practical application of machine learning in embedded systems. She explains the distinction between training and inference, highlighting the challenge of integrating black-box neural network outputs into deterministic embedded software. Malone emphasizes that data science is largely about data preparation and problem formulation, not just algorithm selection, and advises engineers to build projects to overcome imposter syndrome.

    Embedded firmwareAI & machine learning
    10 chapters

  • How I Built This · 34m

    HIBT Lab! Tala: Shivani Siroya

    Shivani Siroya, founder of Tala, explains how her work with the UN Population Fund in Africa and India revealed that micro-entrepreneurs lacked credit histories despite having strong repayment potential. She describes manually lending her own savings to individuals like a tile maker named Seema to prove that daily life data could predict creditworthiness. Siroya details how Tala evolved from this manual process into a fintech app that uses mobile phone data to score and lend to underbanked populations in Kenya, India, the Philippines, and Mexico. The episode covers Tala's business model, its low default rates, and its goal of building a global financial ecosystem for the underserved.

    Starting a product companyValidation & product-market fit
    7 chapters

  • Embedded.fm · 1h 05m

    437: Chirping With the Experts

    Daniel Situnayake, Head of Machine Learning at Edge Impulse and co-author of the O'Reilly book 'AI at the Edge', discusses the practical application of embedded machine learning. He explains the BLERP framework (Bandwidth, Latency, Economics, Reliability, Privacy) for deciding when to deploy AI on edge devices and emphasizes the importance of responsible AI, including hiring ethicists and using diverse datasets. The conversation also covers the current state of large language models like ChatGPT, noting their utility as creative tools with human-in-the-loop oversight rather than autonomous agents.

    AI & machine learningEmbedded firmware
    10 chapters

The summary, chapters and topics are generated from the episode’s audio by this directory, not written by EE Times Current. The episode itself belongs to its publisher.