What is Artificial Intelligence

A Roadmap for Beginners in Machine Learning with many valuable resources for any ML workers or enthusiasts + how to stay up-to-date with news This guide is intended for anyone having zero or a small background in programming, maths, and machine learning. There is no specific order to follow, but a classic path would be from top to bottom. If you don't like reading books, skip it, if you don't want to follow an online course, you can skip it as well. There is not a single way to become a machine learning expert and with motivation, you can absolutely achieve it. The video: https://youtu.be/RirEw-uaS_8?list=PLO4GrDnQanVfb6Ins6up1xScJHl8YuwPQ The complete article: https://ift.tt/SRO2dgq All the links on GitHub: https://ift.tt/lGOZgLi Artificial is a fantastic field, but it goes extremely fast. Don't miss out on the most important and exciting news by joining great communities, people, newsletters, and more you can all find in this guide! posted on Instagram - https://instagr.am/p/Ca4c1WSOEY4/

Impressive restoration of old pictures by AI ! GFPGAN, a new AI model for photo restoration that works well even with very low or very high-quality inputs. The code is available and you can try it on one of the many online apps implementing it. A video about GFPGAN is coming this weekend on my channel, stay tuned! Code: https://ift.tt/mk8HC3r Try it: https://ift.tt/lQKqMfv Paper: https://ift.tt/ydoASRg posted on Instagram - https://instagr.am/p/Ca14yOHulHj/

Throwback to image-GPT from OpenAI: A big step in image synthesis back in late 2020. It can generate the pixels of half of a picture using an NLP model! Learn more about i-GPT (link in story): https://youtu.be/FwXQ568_io0 @gpt3.ai posted on Instagram - https://instagr.am/p/CZbvflfuhlk/

The paper is what’s most interesting. Change my mind. posted on Instagram - https://instagr.am/p/CZRcThMujfi/

Do not underestimate the power of sleep! Learn more in the video short 👇 https://youtu.be/KRB-7HEfTdU @drmattwalker posted on Instagram - https://instagr.am/p/CZMawyuO8oe/

Maybe we should call artificial intelligence “massive data interpolators” and not give false hopes posted on Instagram - https://instagr.am/p/CZHJ1cWuClz/

Generalization is one of the biggest challenges in machine learning due to how it learns: from data. How can we make the model work on any examples after only training on a few (or many but finite) examples? We use many tricks like data augmentation, normalization, etc., but its ability to generalize is still (extremely) far from humans. posted on Instagram - https://instagr.am/p/CY3tqRyL4Li/

This is so true 😂 Can you relate? I think we all lived both situations, and some of us live through the imposter syndrome much more often, which is not desired! I know that this is hard, but please, try to treat yourself like you would treat others! You would never denigrate one of your friends for succeeding, so why are you doing it to yourself? See the positive sides of your successes and take the time to appreciate them, don’t strictly focus on what went wrong! I believe being on the god complex most of the time is just as hurtful, but I don’t know anyone in this state of mind, so I’d assume most people are on a relatively low rollercoaster with some spikes and steep drops to the impostor syndrome. Take care of yourself both physically and mentally! First shared by @hackernoon posted on Instagram - https://instagr.am/p/CYZA0omg_MC/

What Can AI Really Do in 2021? AI Rewind + Highlights ft. @yuval_noah_harari & @kaifu_lee 👉 https://youtu.be/z5slE_akZmc Link in bio! posted on Instagram - https://instagr.am/p/CX6IbeJLMyh/

Of course, this is a harsh simplification and isn’t entirely true, but I like its simplicity and humor in explaining the first three. A tensor is just what acts as a tensor following coordinate change. It can be of any rank or dimension and is used to describe physical properties. A matrix is used to describe a 2D tensor using a coordinate system, but the matrix is not the tensor. Similarly: a row matrix is used to describe a 1D tensor (vector), a number is used to describe a OD tensor (scalar), and a 3D matrix is used to describe a 3D tensor. Likewise for all dimensions. posted on Instagram - https://instagr.am/p/CX3pnw3geiT/

Do you have difficulty keeping up with new research and applications in AI / ML, and your "to-read" list just gets bigger? 🧐 Easy fix: Follow my channel or blog where I explain new applications/papers in short videos and articles weekly! 👇 I hope it can be useful to you and let me know your thoughts on the content so I can improve and better answer this need! Link in bio (https://youtube.com/c/WhatsAI) posted on Instagram - https://instagr.am/p/CWx6ZNVryz1/

We’ve just launched designstripe on Product Hunt! Product Hunt is a curated collection of the best new tech products, every day. Come check us out (TODAY only) and spread the love - any type of support will help tremendously. My team will be around all day to answer your questions and comments. We can’t wait to hear what you think! Quick Designstripe overview: we are building tools for limitless creativity and possibilities. Beautiful, customizable illustrations! Check it out, take a second to sign up, and support us with your comments and feedback on the platform with live answers! 😄 Our company on Product Hunt (link in story and on @designstripeapp ): https://ift.tt/3DFnr9o posted on Instagram - https://instagr.am/p/CWiX8bCLadt/

Can you relate? 😂 posted on Instagram - https://instagr.am/p/CWIupgEAqH3/

2021: A Year Full of Amazing AI papers - A Review [work in progress...] 👇 A curated list of the latest breakthroughs in AI by release date with a clear video explanation, link to a more in-depth article, and code. This is a work in progress... Star this repository to stay up to date! ⭐️ Feel free to message me any interesting paper I may have missed to add to this repository! Link in bio! https://ift.tt/3BTsgKz posted on Instagram - https://instagr.am/p/CV-h5keL1kx/

Just like most things now, there are too many online courses and a lot of them are noise in our dataset. The challenge is to find the great and useful ones, which is why I am always working on improving my curated list of free resources to learn AI, keeping it up to date and relevant for anyone starting in the field. If you want to start or need to improve your data science/AI skills, check it out! Link in bio! https://ift.tt/3CrWndf posted on Instagram - https://instagr.am/p/CV2lb6iLoqb/

My AI Monthly Top 3 - October 2021 is out! Let me know what you guys think of the papers covered in this iteration! Link in bio! https://ift.tt/3CyIgCU posted on Instagram - https://instagr.am/p/CVu4M_DrvQZ/

If you crawl the web, you will absorb human biases... A new study examines the issue of training large language models on the web like GPT-3. Abeba Birhane and colleagues at University College Dublin and University of Edinburgh audited the LAION-400M dataset of paired images and text drawn from Common Crawl, a large repository of web data. Using CLIP, another model trained on the web able to score the similarity between text and images, they could find a high correspondence between words and pictures that are frequently associated with one another on the web, even if the associations are spurious or otherwise undesirable. As you can see in these extreme example images, the authors entered text queries into LAION-400M’s search function, which returned matching images with similarity measured using CLIP and discovered that the dataset absorbed many biases that can be found on the web. As they state: "Due to the gender biases built into CLIP, the odds of the misogynistic one making it through the filtering process might be higher. The main point here is not that we successfully generated provocative examples but that the sheer ease of producing such so-termed ‘‘corner cases” emanates directly from the strong mis-associations baked into the model that can potentially amplify selection bias towards offensive samples in the CC corpus." Read the study: https://ift.tt/3jeEVkG posted on Instagram - https://instagr.am/p/CVMApabALXt/