Hey khazeema saleem Al here

very concise; what an amazing video.

I disagree that people are being judged for being bad at math more harshly than for other subjects. Compare "I'm not interested in history" or "I don't like reading" with "I'm bad at math".

UMAP rocks! The only problem I see is the explainability of this high dimensionality reduction, which is easily done in PCA. In other words, you can get the best variables to explain the clustering, which is important when you are focusing on variable selection. What do you think?

You’re the best 👏👏👏

Thank you very much for this thorough, well-curated, and comprehensive review of MAMBA.

How are exploration and exploitation abilities or forms of intelligence? To me they're more like competing actions, like going to the shops vs going to the cinema. I am still capable of both actions.

sooo well explain, brilliant!

Another great video, as usual! This little bean mutant of yours always puts a smile on my face ☺ Is it possible that it is actually an AI? For example, a transformer that converts language information into the facial expressions of the animated bean. That would be so cool 😎 I have a question: I am looking for training methods that are not based on backpropagation. Specifically, I want to avoid running backwards through the NNW again after the forward pass. Do you know of any algorithms like this? Already 2^10 * Thanks in advance 😄

Thanks!

Neat

You were the only lecturer I was able to pass some machine-learning related courses. now i know why :D you really took huge effort while staying nice/helpful/flexible... all the time

7:05 Spoiler alert! Are you the only person wearing red?

To brew my cup of coffee I need scale with 0.1 gram accuracy and timer to follow best recipe, so much math to achieve one tasty cup of coffee :)

I was tortured and bullied by math in school. I did engineering of automated systems, it was a nighthmare. I hate math because of the messy notation of variables.

Creating AI that's actually good at things that are very hard for humans often seems more important; However, tasks that are the easiest for humans can make great benchmarks for AI. The oldest idea in that direction is of course the Turing Test.

I believe math attacked me once in school. I haven't been the same since.

I don´t fear math, I just can´t follow much of the stuff. Now I´m even more frustrated, cause I couldn´t find you on the foto, ok, spatial thinking is connected with math talent, right? 😉Thanks for sharing your thoughts.

no. obviously you have no idea about the matter and have no clue about psychometrics at all. yes, there is an effect called "subscale scatter" which becomes even more pronounced the higher the iq is you're testing for, but there is such a thing as a general factor to fluid intelligence. this is established fact and no one in their right mind is questioning this.

Maths is one of my biggest fears 😮

Congratulations! Your description of being a lecturer giving you both a decent salary and academic freedom sounds very different from what I've heard about US universities. I enjoy Dr. Angela Collier's channel for her insights into physics and academics. She is also a great science communicator so you have that in common too. :) It would be very interesting for you two to have a conversation comparing your experiences. AI might even come up as a spicy side topic.

Great personal and professionl growth

Wow! Just great ❤

Pleaseeeee do image position embeddings as well! This one was so good!

Thank you so much !! you really super simplified it for any beginner level deep learner to understand

Good explanation. Most videos on attention focus on how it's implemented, not on the design choices behind it. To my understanding, the goal was to mitigate the computational inefficiencies of RNNs and the spatial limitations of CNNs in order to achieve a universal representation of a sequence. I wanted to clarify one thing: you depicted multiple FFNNs similarly to how RNNs are usually rolled out. Is it just the same one FFNN that takes a single attention-encoded vector as input and predicts the next token from this ONE vector? By the way, what brand is that sweater? Loro Piana? :)

Congratulations 🎉

Yes, when a person refers to the human brain in comparison to AI, they generally mean the collective intelligence of humanity, rather than the capabilities of an individual brain.

IT majors in India used to recruit for general intelligence and then make it sparse in the profession, focusing on specialized, repetitive tasks rather than broad skill development.

Interesting. It's almost like two types of tokens: nodes + edges which can each be compressed to a feature vector. But yes, with positional encoding you're left with "random walk with restart" or a traversal depth. Or one could sum node_vector + edge_vector ~= positional distance. but yeah, more graph solutions coming in the future.

very helpful, my thanks

Nice work!

Great video

i am so confused

I really like your vids

Super tare! Succes in continuare

I'm interested even more on the generative side, generating large graphs with contained text in them from a prompt, can be useful for modalities which are represented by large graphs. I've yet to see anyone doing this. While you can prompt LLMs to generate small graphs, for larger graphs you see significant performance drops.

Beautifully explained by a smart & beautiful person!

I need to read this more deeply, I don't understand why would just grafting the parameters willy nilly works

Hello Letitia!!! First off, I love your videos! they are helping me a lot throughout my PhD (which, I must add, takes a lot of inspiration from your research too!). Second, I am currently interested in Continual Learning applied to MMLT, so I'm thinking to study some of these multi modal transformers. What do you think? is this an interesting line of research? PD: congrats on your degree!

Excellent! need to go deeper that could be a game changer for reasoning, as it makes more sense to reason on a graph rather than on the next token.

Thank you for this video Letitia! As always amazing :=) Side note: Loved the silent Good Bye :)))

Its interesting to see "attention" on graph structures again. I think in the future a more structured knowledge representation may play a role on improving reasoning, as we could leverage logic and rules using engines on them, like compilers aid in code generation.

It would be interesting how much computation this needs

Letitia, you are an amazing teacher/instructor!

LOVE IT. Dr any idea on research in this area?

I understand compleatly nothing. I was focused on this guy aspergery passion. How can I work with such cyborgs :D sorry. I am a hater here but I started questioning my role on the job market after watching it.

At 13:14 - Is this 'clip guided diffusion' done by adding a term to the loss function or via a different method?

❤ I am a fan ❤

Do we have an evaluation of the difficulty to learn each human language in terms of flops to achieve a given perplexity ?