
40 segments available
Daphne Koller is a professor of computer science at Stanford University, a co-founder of Coursera with Andrew Ng and Founder and CEO of insitro, a company at the intersection of machine learning and biomedicine. Support this podcast by signing up with these sponsors: - Cash App - use code "LexPodcast" and download: - Cash App (App Store): https://apple.co/2sPrUHe - Cash App (Google Play): https://bit.ly/2MlvP5w EPISODE LINKS: Daphne's Twitter: https://twitter.com/daphnekoller Daphne's Website: https://ai.stanford.edu/users/koller/index.html Insitro: http://insitro.com PODCAST INFO: Podcast website: https://lexfridman.com/podcast Apple Podcasts: https://apple.co/2lwqZIr Spotify: https://spoti.fi/2nEwCF8 RSS: https://lexfridman.com/feed/podcast/ Full episodes playlist: https://www.youtube.com/playlist?list=PLrAXtmErZgOdP_8GztsuKi9nrraNbKKp4 Clips playlist: https://www.youtube.com/playlist?list=PLrAXtmErZgOeciFP3CBCIEElOJeitOr41 OUTLINE: 0:00 - Introduction 2:22 - Will we one day cure all disease? 6:31 - Longevity 10:16 - Role of machine learning in treating diseases 13:05 - A personal journey to medicine 16:25 - Insitro and disease-in-a-dish models 33:25 - What diseases can be helped with disease-in-a-dish approaches? 36:43 - Coursera and education 49:04 - Advice to people interested in AI 50:52 - Beautiful idea in deep learning 55:10 - Uncertainty in AI 58:29 - AGI and AI safety 1:06:52 - Are most people good? 1:09:04 - Meaning of life CONNECT: - Subscribe to this YouTube channel - Twitter: https://twitter.com/lexfridman - LinkedIn: https://www.linkedin.com/in/lexfridman - Facebook: https://www.facebook.com/LexFridmanPage - Instagram: https://www.instagram.com/lexfridman - Medium: https://medium.com/@lexfridman - Support on Patreon: https://www.patreon.com/lexfridman
Daphne Koller, a prominent figure in computer science and biomedicine, introduces herself as a professor at Stanford and co-founder of Coursera. She discusses her current focus on using machine learning to revolutionize drug discovery and treatment development, especially in light of the ongoing pandemic.
"the following is a conversation with Daphne Koller a professor of computer science at Stanford University a co-founder of Coursera with Andrew Eng and founder and CEO of in seat row a company at the i..."
Daphne Koller addresses the philosophical question of whether we will ever cure all major diseases. She emphasizes the complexity of curing diseases, noting that significant damage often occurs before a disease is diagnosed, making complete cures challenging.
"so you co-founded Coursera I made a huge impact in the global education of AI and after five years in August 2016 wrote a blog post saying that you're stepping away and wrote quote it's time for me to..."
Koller discusses the current understanding of major diseases, highlighting that while some diseases are better understood, others like Alzheimer's and schizophrenia remain poorly understood. She emphasizes the need for a deeper understanding of the mechanisms behind these diseases.
"time and I don't like to make predictions of the type we will never be able to do X because I think that's a you know that's the smacks of hubris it seems that never and in in in the entire eternity o..."
Daphne Koller explores the relationship between disease and longevity, noting that the risk of diseases increases with age. She discusses how aging processes contribute to disease and the potential for overlapping mechanisms between aging and various diseases.
"or to the 80 I think Alzheimer's is probably closer to zero than to 80 there are hypotheses but I don't think those hypotheses have as of yet been sufficiently validated that we believe them to be tru..."
Koller expresses her belief in the importance of increasing health span rather than seeking immortality. She advocates for a longer period of healthy living, aiming for a quality of life that allows individuals to remain active and healthy into old age.
"where we're at okay it is a little unfortunate that would get older and it seems that there's some correlation with the fact the the occurrence of diseases or the fact that we'll get all there mm-hmm ..."
Daphne Koller discusses the evolving role of machine learning in healthcare, emphasizing the need for large, high-quality datasets to drive effective predictive models. She highlights the shift towards using machine learning as a primary tool in addressing health challenges.
"goal and anyway in a grand time the age of the universe it's all pretty short so from the perspective you've done obviously a lot of incredible work on machine learning so what role do you think data ..."
Koller shares her personal journey into the intersection of machine learning and health, tracing her interest back to the early 2000s. She reflects on her father's battle with an autoimmune disease and how it fueled her passion for improving drug discovery.
"learning in my opinion can only be really successfully applied especially the more powerful models if you give it data that is of sufficient scale and sufficient quality so how do you create those dat..."
Daphne Koller introduces the concept of 'disease-in-a-dish' models, explaining how they differ from traditional animal models. She discusses the potential of these models to provide better insights into diseases that have been difficult to study effectively.
"had an autoimmune disease that settled in his lungs and the doctors basis it well there was only one thing we could do which is give him prednisone at some point I remember doctor even came and said h..."
In this segment, Koller elaborates on the advantages of disease-in-a-dish models over traditional animal testing. She explains how these models can provide insights into human-specific disease mechanisms and the potential for personalized medicine. Koller emphasizes the role of induced pluripotent stem cells (iPSCs) in creating these models, allowing researchers to study diseases at a cellular level.
"sure so an animal models for disease is where you create effectively its what it sounds like it's it's a oftentimes a mouse where we have introduced some external perturbation that creates the disease..."
Koller explains the groundbreaking process of reverting adult cells to stem cell status, known as the Yamanaka factors. She discusses the implications of this technology for understanding diseases and potential treatments. Koller highlights the challenges and nuances of this process, including the retention of cellular 'memory' from the original cells.
"just because the findings that we had in the mouse don't translate to a human the disease in the dish bottles is a fairly new approach it's been enabled by technologies that have not existed for more ..."
In this segment, Koller addresses the scalability of creating induced pluripotent stem cells (iPSCs) for research. She discusses the current limitations in generating large numbers of iPSCs and the importance of diversity in genetic backgrounds for studying diseases. Koller also highlights the potential of CRISPR technology to introduce specific mutations for comparative studies.
"maybe that will help also the more global clinical phenotypes that's really what we're hoping to do that step that backward step I was reading about it the Yamanaka factor yes so think that the revers..."
Koller discusses the genetic variability among individuals and its impact on disease susceptibility. She explains the concept of polygenic risk scores and how they quantify an individual's risk for certain diseases based on genetic variations. This segment emphasizes the importance of understanding genetic factors in developing targeted therapies.
"huge challenge or or not so the reverse the reversal is not as of this point something that can be done at the scale of tens of thousands or hundreds of thousands I think total number of stem cells or..."
Koller explores how machine learning can be applied to analyze data from disease-in-a-dish models. She discusses the potential of machine learning to uncover patterns in cellular data that correlate with disease phenotypes. This segment highlights the role of data-driven approaches in advancing drug discovery and understanding disease mechanisms.
"because you also want to capture ethnic background and how that affects things but maybe you don't need one from every single patient with every single type of disease because we have other tools at o..."
In this segment, Koller explains the various ways data from disease-in-a-dish models can be utilized in drug discovery. She contrasts analytical methods that work backward from observed phenotypes with forward approaches that test gene perturbations. Koller emphasizes the importance of machine learning in identifying subtypes of diseases and potential interventions.
"so there's definitely a lot that our genetics contributes to disease risk even if it's not by any stretch the full explanation and from the machine learning perspective their signal there there is def..."
In this segment, Koller elaborates on how machine learning can uncover patterns in biological data to aid drug discovery. She explains the analytical methods used to identify disease mechanisms and potential interventions, emphasizing a less hypothesis-driven approach. This discussion reveals the potential for machine learning to revolutionize how we understand and treat diseases, offering hope for meaningful clinical benefits.
"think there's different ways in which this data could potentially be used some people use it for scientific discovery and say oh look we see this phenotype at the cellular level so let's try and work ..."
Koller shares her excitement about the real-world impact of machine learning in biomedicine. She reflects on the motivation of her team at insitro, highlighting their desire to work on projects that benefit humanity. This segment captures the passion driving innovation in the intersection of technology and healthcare, emphasizing the importance of meaningful work in the field.
"and might give rise to things that are not the same things that everyone else is already looking at that's uh I don't know I'm just like to psychoanalyze my own feeling about our discussion currently ..."
Daphne Koller discusses the types of diseases that can be studied using disease-in-a-dish models. She emphasizes the importance of genetic basis and reproducibility in selecting diseases for research. Koller expresses caution about making promises regarding cures but highlights the potential of these models to advance understanding and treatment of diseases like Alzheimer's and type 2 diabetes.
"certainly all of our machine learning people are outstanding and could go get a job you know selling ads online or doing commerce or even self-driving cars yes but but I think they would want they the..."
In this segment, Koller explains the advancements in stem cell-derived models and their implications for studying diseases. She discusses the development of organoids and the potential for multi-organ systems in research. Koller expresses optimism about future capabilities in modeling diseases that are currently challenging to study, showcasing the rapid evolution of biomedicine.
"think that it's really smart bioengineers out there are developing better and better systems all the time so the diseases that might not be tractable today might be tractable in three years so for ins..."
Koller shares the origin story of MOOCs and Coursera, detailing the excitement at Stanford University in the late 2000s about online education. She discusses the collaborative efforts that led to the launch of MOOCs, emphasizing the potential for improved teaching quality and scalability. This segment highlights the transformative impact of online learning on global education.
"and this conversation would seem almost outdated with a kind of scale that could be achieved in like three years that would be so cool the you've co-founded Coursera with injurying and were part of th..."
In this segment, Koller reflects on the global demand for education and the need for continuous learning in a rapidly changing job market. She discusses how traditional education models are becoming less relevant and the importance of providing access to new skills. This conversation underscores the significance of MOOCs in meeting the evolving educational needs of learners worldwide.
"I think the origin story of MOOCs emanates from a number of efforts that occurred at Stanford University around you know the late 2000s where different individuals within Stanford myself included were..."
Koller shares insights gained from her experience with online education, emphasizing the importance of brevity and engagement in course design. She discusses the effectiveness of short video modules and the need for immediate feedback to enhance learner engagement. This segment provides valuable lessons for educators looking to adapt to the needs of modern learners.
"is probably so anyway we so those courses launched in in the fall of 2011 and there were within a matter of weeks with no real publicity campaign just a New York Times article that went viral about a ..."
Daphne Koller discusses the evolution of online learning at Coursera, emphasizing the importance of short video modules. She explains how breaking content into smaller units enhances engagement and allows learners to complete meaningful segments quickly, catering to modern attention spans.
"with short video modules and over time we made them shorter because we realized that 15 minutes was still too long if you want to fit in when you're waiting in line for your kids doctor's appointment ..."
Koller highlights the significance of immediate feedback in online education. She shares insights on how incorporating micro quizzes and self-graded assessments can boost learner engagement and retention, validating her intuition with data-driven results.
"important we learned that engagement during the content is important and the quicker you give people feedback the more likely they are to be engaged hence the introduction of these which we actually w..."
In this segment, Koller reveals findings from experiments on gender bias in online education. She discusses how having female role models as instructors can positively influence the gender balance in STEM courses, showcasing the potential of online platforms for A/B testing.
"exciting but so the shortness the compression I mean that's actually so that that probably is true for all you know good editing is always just compressing the content making it shorter so that puts a..."
Koller reflects on the challenges of teaching effectively in both online and face-to-face settings. She argues that innovative pedagogical approaches require more preparation and effort from instructors, contrasting them with traditional lecture formats.
"what MOOCs do is that they give you these chunks of content and then ask you to practice with it and that's where I think some of the newer pedagogy that people are adopting and face-to-face teaching ..."
Koller shares her perspective on the future of MOOCs and their role in education. She believes that while MOOCs won't replace face-to-face teaching, they will increasingly supplement traditional education, especially for continuing education in a rapidly changing world.
"require a heck of a lot more preparation and so it's one of the challenges I think that people have that we had when trying to convince instructors to teach on Coursera and it's part of the challenges..."
Koller offers advice for those interested in artificial intelligence and machine learning. She emphasizes the importance of building a strong foundation in mathematics and programming before diving into machine learning, encouraging learners to tackle practical problems.
"in the same way that you know recorded music has not replaced live concerts but I do think that especially when you are thinking about continuing education the stuff that people get when they're tradi..."
In this segment, Koller reflects on the most beautiful concepts in deep learning. She discusses end-to-end training and the idea of learning representations that can be reused for different tasks, highlighting their significance in the evolution of AI.
"briefly I know it might be a difficult question to ask but there's a lot of people fascinated by artificial intelligence by machine learning but deep learning is there a recommendation for the next ye..."
Koller shares her surprise at the effectiveness of neural networks in high-dimensional spaces. She discusses the evolution of machine learning and the shift from constrained models to more flexible approaches that leverage large datasets.
"actually start solving problems try and find someone to solve them with because especially at the beginning is useful to have someone to bounce ideas off and fix mistakes that you make and and you can..."
Koller addresses the limitations of machine learning models regarding uncertainty. She emphasizes the need for calibrated predictions, especially in critical applications like medical diagnosis, where incorrect confidence can have dire consequences.
"or interesting perhaps not just deep learning but AI in general statistics good answer with two things one would be the foundational concept of end to end training which is that you start from the raw..."
Koller discusses the feasibility of Artificial General Intelligence (AGI) and its implications. She reflects on the progress made in AI while acknowledging the challenges that remain, particularly in creating systems that can generalize across various domains.
"so end to end and transfer learning yeah is it surprising to you that neural networks are able to in many cases do these things it says it may be taking back to when you when you first would dive deep..."
In this segment, Koller explores the necessity of incorporating insights into AI model architecture. She questions whether a completely knowledge-free approach can lead to effective learning systems, emphasizing the importance of domain-specific knowledge.
"get to a meaningful answer and I think it was true at the time what I think is is still a question is will a completely knowledge free approach where there's no prior knowledge going into the construc..."
Koller elaborates on the importance of calibrating AI predictions to ensure reliability. She discusses various methods, including Bayesian approaches and ensemble techniques, that can help improve the confidence levels of machine learning models.
"with the universal learning machine I don't know I wonder if there's always has to be some insight injected somewhere or whether it can converge so you've done a lot of interesting work with probabili..."
Koller emphasizes the need for human oversight in AI systems, particularly in high-stakes scenarios like autonomous driving. She discusses the importance of ensuring that AI systems can express uncertainty and seek human intervention when necessary.
"when you think for instance about medical diagnosis as being maybe an epitome of how problematic this can be if you were training your network on a certain set of patients on a certain patient populat..."
Koller delves into the philosophical implications of creating intelligent systems. She discusses the balance between advancing AI capabilities and maintaining human control, highlighting the ethical considerations that arise as AI technology evolves.
"that includes the you know medical applications but it also includes you know automated driving because you'd want the network to be able to you know what I have no idea what this blob is that I'm see..."
Koller explores the feasibility of Artificial General Intelligence (AGI), reflecting on its historical aspirations and current limitations. She argues that while machine learning excels in specific domains, it lacks the versatility and adaptability of human intelligence, suggesting that AGI remains a distant goal.
"little bit and then talk about well what controls does one need when protecting when thinking about protections and the AI space so you know I think AGI obviously is a long-standing dream that even ou..."
In this segment, Koller addresses the potential dangers of AI, emphasizing the complexity and unpredictability of current systems. She warns against the risks posed by less intelligent systems and the need for better interpretability and testing to ensure robustness in AI applications.
"of AGI I think that it seems to me a little early today to figure out protections against a human level or superhuman level intelligence who's where we don't even see the skeleton of what that would l..."
Koller discusses the dual-use nature of powerful technologies like machine learning and gene editing. She highlights the potential for misuse and the historical context of technological advancements leading to both beneficial and harmful outcomes, stressing the importance of ethical considerations in their development.
"them and we should and I think it's really important that people are thinking about ways in which we can have better interpret ability of systems better tests for for instance measuring the extent to ..."
Koller shares her optimistic view on human nature, suggesting that most people mean well. She emphasizes the importance of fostering social norms that encourage positive behavior and warns against societal trends that could lead to moral decay.
"technology yeah and on the case of machine learning is at the stereo machine learning so all the kinds of attacks like security almost threats and there's a social engineering with machine learning al..."
In a reflective moment, Koller articulates her personal philosophy on life’s purpose, inspired by Steve Jobs' quote about making a dent in the universe. She emphasizes the responsibility of those born into privilege to contribute positively to society and leave the world better than they found it.
"think you're amazing it's fun to ask a world-class computer scientist and engineer a ridiculously philosophical question like what is the meaning of life let me ask what gives your life meaning what a..."