Deep Learning: Advice on Getting Started with fast.ai - Jeremy Howard | AI Podcast Clips

Duration: 5:00 810 transcript words EN summary

Video summary

Jeremy Howard emphasizes that the most effective way to learn deep learning is through extensive hands-on practice, specifically by training numerous models rather than merely running inference on pre-existing ones. He argues that relying solely on someone else's model prevents learners from developing an intuitive understanding of how inputs affect outputs within their specific domain area. To truly master the field, students must engage in active experimentation: printing out both input and output data, slightly modifying inputs to observe variations in results, and running a wide array of experiments. This approach is central to his course at fast.ai, which recently won the CogX award as the best AI course globally, distinguishing itself by focusing on practical application over theoretical abstraction. A key component of this methodology involves creating custom datasets from scratch using tools provided within the curriculum. Howard illustrates this with an example where a student built a web application capable of differentiating between teddy bears and grizzly or brown bears with nearly 100% accuracy in just four minutes by scraping images via Google Search scripts included in their notebooks. The course provides graphical widgets to assist users in cleaning data, studying results, and identifying specific errors. This practical engagement has led to over a thousand student projects shared on the "Share Your Work" thread, ranging from high-impact academic breakthroughs that outperform state-of-the-art papers to creative applications like classifying hummingbirds native to Trinidad and Tobago. Howard stresses that all of this educational content is offered completely free as a service to the community without any revenue sources. He believes that once individuals grasp the basics and begin training models on their own data, they can become experts in niche areas where society desperately needs specialized knowledge. Instead of producing more generalist researchers, he advocates for cultivating domain-specific expertise using deep learning tools to solve real-world problems such as diagnosing malaria, analyzing language bias, or identifying problem areas in fisheries and ocean conservation. By combining technical proficiency with a passion for a specific field, practitioners can achieve results that surpass those who lack this dual focus on technology and subject matter knowledge. The ultimate goal of this approach is to ensure that research addresses actual problems rather than pursuing abstract innovations without practical utility. Howard contends that the vast majority of interesting research stems from attempting to solve genuine issues effectively, which requires a deep understanding of both the data and the problem context. If one does not work on a real problem they understand, it becomes impossible to determine if results are good or bad, nor can one diagnose why errors occur. Therefore, he concludes that meaningful innovation in areas like transfer learning or active learning still necessitates finding a dataset and domain area that genuinely matters to the researcher, ensuring their efforts contribute tangible value rather than just theoretical advancement.

Read the full video transcript

[Music] so what advice do you have for someone who wants to get started in deep learning train lots of models that's that's how you that's how you learn it so like so I would you know I think it's not just me I think I think our course is very good but also lots of people independently is that it's very good it recently won the cog X award for AI courses as being the best in the world so let's say come to our course cost up faster day I and the thing I keep on harping on in my lessons is train models print out the inputs to the models print out to the outputs to the models like study you know change change the inputs a bit look at how the outputs vary just run lots of experiments to get a you know an intuitive understanding of what's going on to get hooked do you think you mentioned training do you think just running the models inference like if we talk about getting started no you've got to find true in the models so that's that's that's the critical thing because at that point you now have a model that's in your domain area so there's there's there's no point running somebody else's model because it's not your model like so it only takes five minutes to fine-tune a model for the data you care about and in lesson two of the course we teach you how to create your own data set from scratch by scripting google image search so and we show you how to actually create a web application running online so I create one in the course that differentiates between a teddy bear or grizzly bear and a brown bear and it does it with basically hundred percent accuracy took me about four minutes to scrape the images from Google search in the script there's a little graphical widgets we have in the notebook that help you clean up the data set there's other widgets that help you study the results to see where the errors are happening and so now we've got over a thousand replies in our share your work here thread of students saying here's the thing I built and so those people who like and a lot of them are state of the art like somebody said oh I tried looking at different gary characters couldn't believe it the thing that came out was more accurate than the best academic paper after Lesson one and then there's others which are just more kind of fun like somebody who's doing Trinidad and Tobago hummingbirds she said that's kind of their national bird and she's got something that can now classify Trinidad and Tobago hummingbirds so yeah train models fine-tune models with your data set and then study their inputs and outputs how much is fast their courses free everything we do is free we have no revenue sources of any kind it's just a service to the community you're a saint okay once the person understands the basics trains a bunch of models if we look at the scale of years what advice do you have for someone wanting to eventually become an expert train lots of models train lots of models in your domain area so an expert what right we don't need more expert like create slightly evolutionary research in areas that everybody's studying we need experts at using deep learning to diagnose malaria well we need experts at using deep learning to analyze language to study media bias so we need experts in analyzing fisheries to identify problem areas and you know the ocean you know that that's that's what we need so like become the expert in your passion area and this is a tool which you can use just about anything and you'll be able to do that thing better than other people particularly by combining it with your passion and domain expertise so that's really interesting even if you do want to innovate on transfer learning or active learning your thought is it means one i I certainly share is you also need to find a domain or data set that you actually really care for right if you're not working on a real problem that you understand how do you know if you're doing it any good you know how do you know if your results so good how do you know if you're getting bad results why you're getting bad results is it a problem with it like how do you know you're doing anything useful yeah the only to me the only really interesting research is not the only but the vast majority of interesting research is like try and solve an actual problem and solve it really well you