MIT 6.S191: Building AI Models in the Wild

MIT Introduction to Deep Learning 6.S191: Lecture 8
Endless Experimentation: Building AI Models in the Wild
Lecturer: Niko Laskaris (VP, Customer Engineering) and Doug Blank (Head of Research)
Comet ML

Lecture Outline - coming soon!

Subscribe to stay up to date with new deep learning lectures at MIT, or follow us @MITDeepLearning on Twitter and Instagram to stay fully-connected!!

[isatousarr]

Building AI models in the wild refers to the process of deploying and training artificial intelligence models in real-world, unstructured environments where data may be noisy, incomplete, or constantly changing. Unlike controlled laboratory settings where data is curated and processes are streamlined, "in the wild" implies a much more dynamic and unpredictable scenario. This often involves gathering data from diverse sources like sensor networks, social media platforms, IoT devices, or human interactions, and deploying models that must adapt to these conditions in real-time.

Building AI models in the wild presents unique challenges, including data privacy concerns, dealing with biases inherent in real-world data, and ensuring model robustness in diverse and unpredictable environments. For example, AI models trained on data from one geographic location or demographic may perform poorly when applied elsewhere, highlighting the importance of generalization and fairness. Furthermore, models must be designed to handle noisy, missing, or inconsistent data while still making accurate predictions or decisions.

Despite these challenges, there are significant opportunities in deploying AI in real-world settings. From autonomous vehicles navigating busy streets to AI-powered healthcare solutions that assist in diagnosing conditions from diverse patient populations, real-world AI models are enabling transformative applications. Key to success in building AI models in the wild is the continuous feedback loop, where models are updated and retrained based on new data and experiences.

Ultimately, the goal is to build AI systems that are both resilient and adaptable, capable of learning and improving in real-time while being mindful of ethical considerations such as privacy, fairness, and transparency. As AI technology evolves, building models in the wild will continue to push the boundaries of what is possible, driving innovation in fields such as transportation, healthcare, and urban planning.

[UnderstandingCode]

25:52 the description of output bias due to iterative dependency is on point

[htoorutube]

Excellent content! Love it! Still not clear on how prompts work with the LLM, such as system prompts and user prompts etc.

[ra_XOr]

When have this lecture took place? The discussion about ChatGPT feels like a year ago performance! Things are completely different now, at least that what O see and believe.

[terryliu]

I found fascinating that someone asked the question regarding encouraging gpt to ask questions. Think about how human beings learn. Beyond reading lots of info/materials, asking questions might be a great idea of adding some new AI capabilities such as reasoning…

[ghazi]

from 773K views on the 1st lecture to about 24K to last, very few completed!

[HenryNouwen-pdox]

The presenters (particularly Doug B) refers to Doug Eck. Did I miss a talk by Doug Eck?

[PAAOMahesh]

Thanks for your work and looking for more videos on ai

[everything_news]

Will you guys post the last lecture i.e.: Project Presentations here on YT?

[akazawashunichi]

Is this Lecture 7 or 8? If 8, would Lecture 7 become available?

[akshatchouhan]

It was a nice lecture. I really enjoyed it.

[mind-blowing_tumbleweed]

Yep, it's straight up dissapointing when you learn how it works. Coin flipper run on insanely expensive hardware.

[reedschrichte]

Please correct me if 'm wrong, but we're working here with the statistical relationships of letters in a word, words in a sentence, sentences in a paragraph, paragraphs in an article or essay, expressed mathematically as probabilities, generated over the available corpus of lexical output, predominantly human.
In the end, it's just a number :-) kindly forgive the oversimplification but I just had to say that.

[andrewign]

Evelyn Hart was indeed a Canadian ballerina.

[reedschrichte]

I hope you can help Google and Amazon they keep sending me ads for products that I already bought wtf seriously and air tickets for places I already went to could you possibly be more useless it would be better to send me random destinations

[yassinee]

"4. LLM results may be racist, unethical, demeaning, and weird" Is Trump an AI?

[thefamousdjx]

I hate lessons were the lecturer keeps asking students what they think almost like they have to read his mind. Its usually the guys that read ahead or work in the industry already that are always answering, making the beginners feel bad and dumb

[ericgonzales]

Im sorry but I think that LLM's are overrated and over-focused on. I want to make a model for trading stocks and cryptos using neural networks. Why is there never a lecture on that, im sure its way more motivating and relevant to everyday people than learning all the history of the world or asking a LLM some questions.