Life with AI: Recent Episodes
Filipe Lauar
In this podcast I explain some hard concepts of AI in a way that anyone can understand. I also show how AI is influencing our lives and we don’t know.
Hey guys, in this episode I try to explain my ideas on why Chain of Thought works. Of course nothing that I say is proved, the idea is really to give my intuitions on it!
Instagram of the podcast: https://www.instagram.com/podcast.lifewithai
Linkedin of the podcast: https://www.linkedin.com/company/life-with-ai
Hey guys, in this episode I talk about the new model of OpenAI, the OpenAI o1, the model that thinks and reflects before answering.
In the episode I share my opinion and my thoughts after the first contact with the model.
Instagram of the podcast: https://www.instagram.com/podcast.lifewithai
Linkedin of the podcast: https://www.linkedin.com/company/life-with-ai
Hey guys, in this episode I talk about the different types of AI. I don't go deep into the models, I focus more on explaining the different types and which kind of problems they solve.
As types of AI I talk about statistics, machine learning, reinforcement learning, deep learning and generative AI.
Instagram of the podcast: https://www.instagram.com/podcast.lifewithai
Linkedin of the podcast: https://www.linkedin.com/company/life-with-ai
Hey guys, this is the last episode of the Llama3 paper. In this episode, I talk about their discussion on the contamination analysis of the benchmarks, the vision and the speech parts of the model.
I hope you enjoyed the series of episodes about the paper, I personally learned a lot!
Instagram of the podcast: https://www.instagram.com/podcast.lifewithai
Linkedin of the podcast: https://www.linkedin.com/company/life-with-ai
Llama 3 paper: https://scontent-cdg4-3.xx.fbcdn.net/v/t39.2365-6/452387774_1036916434819166_4173978747091533306_n.pdf?_nc_cat=104&ccb=1-7&_nc_sid=3c67a6&_nc_ohc=t6egZJ8QdI4Q7kNvgEUMsBZ&_nc_ht=scontent-cdg4-3.xx&oh=00_AYBeifNn3pUDhDb136i_WQ_jpoYwLgNExZHcNvDV-N1rRA&oe=66A804CD
Hey guys, in this episode I talk about the Llama 3 paper pre-training and post-training!Instagram of the podcast: https://www.instagram.com/podcast.lifewithaiLinkedin of the podcast: https://www.linkedin.com/company/life-with-ai
Hey guys, FINALLY we have Llama 3 paper, the release of the 405B model and the update of the 8B and 70B models. In this episode I give my thoughts about the paper and also an overview about it. In the next episode I will go more deeper in more details of the paper.
Instagram of the podcast: https://www.instagram.com/podcast.lifewithai
Linkedin of the podcast: https://www.linkedin.com/company/life-with-ai
Llama 3 paper: https://scontent-cdg4-3.xx.fbcdn.net/v/t39.2365-6/452387774_1036916434819166_4173978747091533306_n.pdf?_nc_cat=104&ccb=1-7&_nc_sid=3c67a6&_nc_ohc=t6egZJ8QdI4Q7kNvgEUMsBZ&_nc_ht=scontent-cdg4-3.xx&oh=00_AYBeifNn3pUDhDb136i_WQ_jpoYwLgNExZHcNvDV-N1rRA&oe=66A804CD
Hey guys, in this episode I talk about how to choose the best model for your AI application. I discuss the different tradeoffs you should take into considering before choosing which model to test and to deploy.
Transformers tutorials github by Niels Rogge: https://github.com/NielsRogge/Transformers-Tutorials
Instagram of the podcast: https://www.instagram.com/podcast.lifewithai
Linkedin of the podcast: https://www.linkedin.com/company/life-with-ai
Hey guys, in this episode of the podcast I talk about Stable Diffusion, a famous open source image generation algorithm.
Instagram of the podcast: https://www.instagram.com/podcast.lifewithai
Linkedin of the podcast: https://www.linkedin.com/company/life-with-ai
Stable diffusion paper: https://arxiv.org/pdf/2112.10752
Good Medium post about SD: https://medium.com/@steinsfu/stable-diffusion-clearly-explained-ed008044e07e#97f4
Amazing YouTube videos explaining diffusion: https://www.youtube.com/watch?v=1CIpzeNxIhU
Explaining stable diffusion with code: https://www.youtube.com/watch?v=-lz30by8-sU
Hey guys, in this episode I talk about diffusion models, the algorithm behind all the image generation models today, like stable diffusion, Midjourney and dall-e. In the episode I explain the diffusion process along with some other technical concepts that are important, like white noise and markovian process.
Instagram of the podcast: https://www.instagram.com/podcast.lifewithai
Linkedin of the podcast: https://www.linkedin.com/company/life-with-ai
Blog post explaining diffusion: https://www.assemblyai.com/blog/diffusion-models-for-machine-learning-introduction/
Diffusion paper: https://arxiv.org/pdf/2006.11239.pdf?ref=assemblyai.com
Hey guys, in this episode I talk about 3 very important models that use contrastive learning, CLIP, SigLIP and JinaCLIP. They are image-text embedding models that allow us to for instance do image-text retrieval.
CLIP paper: https://arxiv.org/pdf/2103.00020
SigLIP paper: https://arxiv.org/pdf/2303.15343
JinaCLIP paper: https://arxiv.org/pdf/2405.20204
Github of similarities and contrastive loss: https://github.com/filipelauar/projects/blob/main/similarities_and_contrastive_loss.ipynb
Instagram of the podcast: https://www.instagram.com/podcast.lifewithai
Linkedin of the podcast: https://www.linkedin.com/company/life-with-ai
Hey guys, in this episode I talk about two very important technical concepts in Deep Learning, constrastive learning and cosine similarity. They are very useful when training embedding models or doing RAG.
Very good blog post about contrastive losses: https://lilianweng.github.io/posts/2021-05-31-contrastive/
SimCLR paper: https://arxiv.org/abs/2002.05709
Instagram of the podcast: https://www.instagram.com/podcast.lifewithai
Linkedin of the podcast: https://www.linkedin.com/company/life-with-ai
Hey guys, in this episode I talk about the FineWeb dataset, the best pre-training open source dataset to date. In the episode I explain how they created the dataset and I also share some results.
Link to the huggingface blog: https://huggingface.co/spaces/HuggingFaceFW/blogpost-fineweb-v1
Instagram of the podcast: https://www.instagram.com/podcast.lifewithai
Linkedin of the podcast: https://www.linkedin.com/company/life-with-ai
Fala galera, nesse episódio eu falo com o João Batista, Technical Product Manager da Stackspot AI. No episódio a gente falou bastante sobre o uso de LLMs como um copilot usando os próprios documentos da empresa para dar as respostas usando RAG.
Hey guys, in the brazilian version of the podcast I discussed with Joao from Stackspot AI. In the episode I talk about how they are developing enterprise copilot assistants using RAG.
In the episode we talk both about technical and product aspects, like similarity metrics, how many documents to use, how to show the answer to the user, how to metrify the quality of the answers...
Linkedin do Joao: https://www.linkedin.com/in/joaobatista-cordeironeto/
Linkedin da Stackspot AI: https://www.linkedin.com/company/stackspot/
Instagram do podcast: https://www.instagram.com/podcast.lifewithai
Linkedin do podcast: https://www.linkedin.com/company/life-with-ai
Hey guys, in this episode I talk about two papers, BitNet and 1.58 bit Transformer. These two papers from microsoft tell a new receipe to train 1 bit transformers, improve hugely the memory and energy consumption along with lower inference times.
BitNet paper: https://arxiv.org/pdf/2310.11453
1.58 bit paper: https://arxiv.org/pdf/2402.17764
Instagram of the podcast: https://www.instagram.com/podcast.lifewithai
Linkedin of the podcast: https://www.linkedin.com/company/life-with-ai
Hey guys, continuing the series of episodes about PEFT, in this episode I talk about inference optimization techniques for LLMs.
I talk about layer pruning, where we prune consecutive layers of the LLM without almost not losing model performance.
I also talk about Mixture of Depths, a similar technique to Mixture of Experts, where we have a router that choses which tokens will be processed in which layer of the LLM.
Paper MoD: https://arxiv.org/pdf/2404.02258.pdf
Paper layer pruning: https://arxiv.org/pdf/2403.17887v1.pdf
Instagram of the podcast: https://www.instagram.com/podcast.lifewithai
Linkedin of the podcast: https://www.linkedin.com/company/life-with-ai
Hey guys, this is the first episode in a series of episodes about PEFT, Parameter Efficient Fine Tuning. In this episode I talk about LoRA and QLoRA, two widely used methods that allowed us to fine tune LLMs way faster and in a single GPU without losing performance.
Video sobre QLoRA: https://www.youtube.com/watch?v=6l8GZDPbFn8
LoRA paper: https://arxiv.org/pdf/2106.09685.pdf
QLoRA paper: https://arxiv.org/pdf/2305.14314.pdf
Instagram do podcast: https://www.instagram.com/podcast.lifewithai
Linkedin do podcast: https://www.linkedin.com/company/life-with-ai
Hello, in this episode I talk a Retrieval Aware Fine Tuning (RAFT), a paper that proposes a new technique to use both domain specific fine-tuning and RAG to improve the retrieval capabilities of LLMs.
In the episode I also talk about another paper that is called RAFT, but this time Reward rAnking Fine Tuning, which proposes a new technique to perform RLHF without the convergence problems of Reinforcement Learning.
Retrieval Aware Fine Tuning: https://arxiv.org/abs/2403.10131v1
Reward rAnking Fine Tuning: https://arxiv.org/pdf/2304.06767.pdf
Instagram of the podcast: https://www.instagram.com/podcast.lifewithai
Linkedin of the podcast: https://www.linkedin.com/company/life-with-ai
Hello guys, in this episode I explain how we can scale the context window of an LLM to more than 1M tokens using Ring Attention. In the episode, I also discuss if RAG is dead or not based on these advancements in the context window.
Paper Lost in the Middle: https://arxiv.org/pdf/2307.03172.pdf
Gemini technical report: https://storage.googleapis.com/deepmind-media/gemini/gemini_v1_5_report.pdf
Paper Ring Attention: https://arxiv.org/pdf/2310.01889.pdf
Instagram of the podcast: https://www.instagram.com/podcast.lifewithai
Linkedin of the podcast: https://www.linkedin.com/company/life-with-ai
Hey guys, in the Brazilian version of the Podcast I interviewed Andre, he is an AI expert on IBM and we talked a lot about how to solve problems using AI.
Brains website: https://brains.dev/
Andre's Linkedin: https://www.linkedin.com/in/andrefelipelopes/
Brains' Linkedin: https://www.linkedin.com/company/brains-brazilian-ai-networks/
Instagram of the podcast: https://www.instagram.com/podcast.lifewithai
Linkedin of the podcast: https://www.linkedin.com/company/life-with-ai
Hey guys, in this episode I talk about Mixture Of Experts, more specifically about Mixtral, which is today the best open-source LLM available, and also better than ChatGPT 3.5 and Gemini Pro.
Mixtral paper: https://arxiv.org/pdf/2401.04088.pdf
Mixtral model: https://huggingface.co/mistralai/Mixtral-8x7B-v0.1
Mixtral YouTube: https://www.youtube.com/watch?v=mwO6v4BlgZQ
Instagram: https://www.instagram.com/podcast.lifewithai
Linkedin: https://www.linkedin.com/company/life-with-ai
Hey guys, in this episode I have Sergei as guest and we talked a lot about NLP and Named Entity Recognition (NER). Sergei and his colleagues at NuMind have the current state or the art model for NER and we discussed a lot about it during the episode.
English model: https://huggingface.co/numind/generic-entity_recognition_NER-v1
Multi-language model: https://huggingface.co/numind/generic-entity_recognition_NER-multilingual-v1
Sergei's Linkedin: https://www.linkedin.com/in/svbogdanov/
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Hey guys, in this episode I explain most of what I know about Transformers. I talk about the architecture, the attention formula, encoder, decoder, self-supervised learning, positional encoding, tokenization, inductive bias, Vision-Transformers, receptive fields...
It was the most technical episode I've recorded so far, and I hope you like it! By the way, it worth listening to this episode with the Transformers paper.
Paper Transformers: https://arxiv.org/pdf/1706.03762.pdf
Link of OpenAI explaining Next Token Prediction: https://www.linkedin.com/posts/zainhas_the-most-clearest-and-crisp-explanation-ugcPost-7132561633280692224-63AX?utm_source=share&utm_medium=member_desktop
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Hey guys, in this episode I explain RAG (Retrieval Augmented Generation) and the concept of agents executing different tasks. Hope you like it!
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Hey guys, in the brazilian version of the podcast I interviewed Daniel, CTO of WeClever, a company that uses AI to improve chat bot experience. In the episode we talked about fine tuning ChatGPT, LoRA, RAG and more!
WeClever: https://www.linkedin.com/company/wecleverco/
Daniel Linkedin: https://www.linkedin.com/in/dmerlimorais/ Instagram: https://www.instagram.com/podcast.lifewithai/ Linkedin: https://www.linkedin.com/company/life-with-ai
Hey guys, this episode was really great! I have tips for both technical and behavioral interviews.
In the technical interview part, I talk about what is always present in interviews and is always good to know! I also talk about the biggest technical mistakes that people make in interviews and explain them. In the behavioral part, I talk a little about the interviewer's vision and what is important to know!
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Hey guys, in this episode I talk about how AI algorithms are trained using supervised and self-supervised learning, how text tokenization works, how ChatGPT was trained and I also talked about document intelligence.
This was a heavy technical content episode and I hope you enjoy it!
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Hey guys, in this episode I talk about how to better use ChatGPT. In the episode I talk about Chain of Thought, zero shot, few shot and more!
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Hello guys, in this episode I go through the leaked document about GPT4 explaining the different points of the architecture, training, inference and dataset. It's a very nice, curious and a bit technical episode!
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Hello everyone, in the brazilian version of podcast I interviewed Juliemar from Amicci and we discussed about private label products, marketplace and how to leverage OpenAI API to build their own products. They a lot of very nice applications of AI at Amicci and Juliemar talked more about them during the episode.
Juliemar's Linkedin: https://www.linkedin.com/in/juliemarberri/
Amicci's website (only in Portuguese): https://amicci.com.br/
Podcast LinkedIn: https://www.linkedin.com/company/life-with-ai
Podcast Instagram: https://www.instagram.com/podcast.lifewithai/
Hi guys, in this episode in the Brazilian podcast I have as guest Vinicius, CEO of MadeinWeb. In the episode we discussed about the use of AI in agriculture and the power of cloud for digital transformation.
Vinicius LinkedIn: https://www.linkedin.com/in/vgallafrio/
MadeinWeb LinkedIn: https://www.linkedin.com/company/madeinweb-mobile/
Podcast LinkedIn: https://www.linkedin.com/company/life-with-ai
Podcast Instagram: https://www.instagram.com/podcast.lifewithai/
Hi everyone, in the portuguese version of the podcast I received as guest Rafael Lanna, CRO of Ewally, a Brazilian fintech for B2B and B2C. In the episode we discussed a lot about how to provide credit score for people that don't have historical transactions.
Rafael's LinkedIn: https://www.linkedin.com/in/rafaellanna/Ewally's website: https://www.ewally.com.br/
Podcast LinkedIn: https://www.linkedin.com/company/life-with-ai
Podcast Instagram: https://www.instagram.com/podcast.lifewithai/
In this episode I have Krish Ramineni, Co-Founder and CEO of Fireflies a meeting recording tool that transcribes your meeting and has search engine with AI superpowers to enable you to keep track on what was discussed. Their engine allows you to perform different tasks like summarisation, keyword search, topic search...
Fireflies is launching soon Fred, their chatGPT like algorithm that will be able to answer everything you want from your meetings. During the episode we discussed a lot about Generative AI, chatGPT and different applications that we will have soon powered by them.
Fireflies website: https://fireflies.ai
Krish's Linkedin: https://www.linkedin.com/in/krishramineni/
Fireflies' Linkedin: https://www.linkedin.com/company/fireflies-inc/
Podcast Linkedin: https://www.linkedin.com/company/life-with-ai
Podcast instagram: https://www.instagram.com/podcast.lifewithai/
Hey guys, in this episode I have Eric Olson, Co-Founder and CEO of Consensus, an evidence-based search engine. In the episode of discussed a lot the technical aspects of building a search engine, going through the different steps of it, like the keyword matching, vector similarity search and also Large Language Models for Q&A. We also discussed about his entrepreneur life, going from a technical position as a data scientist to a CEO position, his new challenges and also his day-to-day work.
Consensus search engine: https://consensus.app/
Eric's Linkedin: https://www.linkedin.com/in/eric-olson-1822a7a6/
Consensus' Linkedin: https://www.linkedin.com/company/consensus-nlp/
Podcast Linkedin: https://www.linkedin.com/company/life-with-ai
Podcast instagram: https://www.instagram.com/podcast.lifewithai/
Hello everyone, in this episode I explain the famous algorithm ChatGPT. ChatGPT is a chatbot developed by OpenAI that is able the answer almost every question. It can be free questions, scientific question or even coding questions. ChatGPT uses GPT3 as backbone and also supervised training along with reinforcement learning using PPO algorithm.ChatGPT: https://chat.openai.comInstagram: https://www.instagram.com/podcast.lifewithai/Linkedin: https://www.linkedin.com/company/life-with-ai
Hello everyone, in this episode I explain how tokenizers work. They are basically what enables us to input the text into a NLP algorithm like BERT or GPT. In the episode I explain 3 types of tokenizers, word based, character based and sub-word based representation.
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Huuging Face blog about tokenizers: https://huggingface.co/docs/transformers/tokenizer_summary
Hey guys, in this episode of the podcast I received Guillaume, COO of Waalaxy. Waalaxy is a CRM focused on prospecting any type of clients on LinkdIn, in the episode, besides explaining the product, we discussed possible use cases of AI in the product e how did they manage to grow from zero to more the 6 millions in Anual Recurrent Revenue and hire and maintain talents specially in the tech field.
Link to get 2 months for free on Waalaxy: https://waal.ink/Zgh3CR Linkedin of Guillaume: https://www.linkedin.com/in/guillaume-portalier-waalaxy/ Instagram of the podcast: https://www.instagram.com/podcast.lifewithai/ Linkedin of the podcast: https://www.linkedin.com/company/life-with-ai
Hey guys, in this episode I talk about the document intelligence algorithms that we have at Qantev, so basically what I do in my job! We have two main document intelligence algorithm, which are information extraction, where we want to retrieve some specific information from a document, and also table extraction, where we want to extract a table in a document into a CSV format.
Instagram of the podcast: https://www.instagram.com/podcast.lifewithai/
Linkedin of the podcast: https://www.linkedin.com/company/life-with-ai
Linkedin of Qantev: https://www.linkedin.com/company/qantev/
Hey guys, this is a very special episode because I received as guest the French Grand Master and actual blitz world champion Maxime Vachier-Lagrave, the famous MVL. In the episode we discuss about the use of engines in chess, about his history and of course about chess in general. This episode is super nice and I hope you all enjoy it!
MVL instagram: https://www.instagram.com/mvl_chess/?hl=en
MVL twitter: https://twitter.com/vachier_lagrave
Instagram of the podcast: https://www.instagram.com/podcast.lifewithai/
Linkedin of the podcast: https://www.linkedin.com/company/life-with-ai
In the episode I interview Hadrie, the CTO at Qantev (company that I work!). In the episode we talked about the company, we explained some parts of our product that we use machine learning and also how is the life of a CTO. We talked about how we detect fraud, waste and abuse in health insurance claims, how we retrieve information from the unstructured data that we have and how we managed the hospitals network of an insurer.
We are just finalizing a fund raising Series A round and we have a lot of open positions for Machine Learning, Data Science, DevOps, infrastructure, full stack (front and back), sales and also product. We can also sponsor visa, so if you want to live in Paris and work with us this can be a great opportunity!
Linkedin of Hadrien:https://www.linkedin.com/in/hadrien-de-march/ Linkedin of Qantev: https://www.linkedin.com/company/qantev/mycompany/ My Linkedin: https://www.linkedin.com/in/filipe-lauar/ Instagram of the podcast: https://www.instagram.com/podcast.lifewithai/ Linkedin of the podcast: https://www.linkedin.com/company/life-with-ai
In this episode I receive as guest my professor at Sorbonne, Jean-Daniel Kant. Jean's research aims to simulate complex systems like the labor market using a model-agent approach, where he models the system and agents behaviours and then analyse the results to understand the consequences of chances in a part of the system, like an introduction of a new law. Jean developed two great simulators, WorkSim and NewJobs and he explains both of them in the episode and also gives some spoilers of his new project TerraNeon, that aims to simulate climate changes. In the end we also talk a little bit about Metaverse.
Jean on Linkedin: https://www.linkedin.com/in/jean-daniel-kant-569937a/
Jean's website: https://webia.lip6.fr/~kant/
Instagram of the podcast:: https://www.instagram.com/podcast.lifewithai/
Linkedin of the podcast: https://www.linkedin.com/company/life-with-ai
In this episode I receive as guest Slater Victoroff, CTO of Indico Data. In the episode we had a great conversation about Artificial Intelligence, we talked about unstructured data, document intelligence, multimodal learning, how humans learn and even general AI! Besides that, Slater also has experience in the Venture Capital world, so we discussed the actual VC market and also what he looks to invest in a startup. The episode is really nice, I learned a lot and really enjoyed the conversation with Slater, I hope you also like it!
Slater on Linkedin: https://www.linkedin.com/in/slatervictoroff/
Slater on Twitter: https://twitter.com/sl8rv
Slater website: https://www.slater.website/
Indico Data: https://indicodata.ai/
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Sparse transformers: https://arxiv.org/abs/1904.10509
Unifying Language Learning Paradigms: https://arxiv.org/abs/2205.05131
The illustrated Transformer: https://jalammar.github.io/illustrated-transformer/
Enso open source library: https://github.com/IndicoDataSolutions/Enso
Follow the information blog post: https://venturebeat.com/2022/06/04/follow-the-information-the-real-data-problem/
Hey guys, in this special episode of 1 year of the podcast a receive my best international friends to talk about our experience of living, studying and working abroad. I have as guests for this special episode Chiara and Matteo from Italy, Zahra from Iran, Santiago from Colombia and Anas from Morocco.
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Hey guys, in this episode I discuss if Deep Learning is for everybody. In the episode I discuss about the availability of deep learning to people to use for both text and image applications showing there diferencies and how the cloud platforms are impacting on it.
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Huuging Face: https://huggingface.co/
Hey guys, in this episode I talk about Gato, a new DeepMind algorithm that is capable of solving more than 600 completely different tasks. This algorithm is for sure a first step to achieve general artificial intelligence, because with only 1 algorithm with 1 set of parameters is able to solve tasks like play games like Atari, caption images, chat with humans (chatbot) e even control a robotic arm to move some blocks.
Paper: https://arxiv.org/pdf/2205.06175.pdf
DeepMind's blog post about the paper: https://www.deepmind.com/publications/a-generalist-agent
Episode about Reinforcement Learning: https://open.spotify.com/episode/2w088ECqtC0CVuAe9l25A2?si=886d21cbbc8f4757
Episode about Transformers: https://open.spotify.com/episode/5qdXu3Y2aq2z6EqMGoM5YW?si=6e8d1139c04c4338
Episode about Vision Transformers: https://open.spotify.com/episode/7xcdNNkvLKjq8QBozluATo?si=952a89557afe4be8
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Hey guys, in this episode I finish a series of 3 episodes about Transformers talking about BERT, the model revolutionized the NLP field. If you want to better understand the architecture of the model and mainly how it was trained, go listen to this episode cause it is really nice!
Artigo do BERT: https://arxiv.org/pdf/1810.04805.pdf
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Hey guys, in this episode I talk about Vision Transformers, the application of Transformers for computer vision (images). In the episode I explain the Vision Transformers architecture and also the idea of applying the technique that was firstly proposed for NLP to images and also its challenges. The episode is really nice and you will understand why Transformers took so long to be applied to images and the future of the computer vision field.
Vision Transformer original paper (ViT): https://arxiv.org/pdf/2010.11929.pdf
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Hey guys, in this episode I finally explain the Transformers network architecture! The paper Attention is all you need proposed the Transformer network and it was groundbreaking for firstly NLP field and now for all the Deep Learning fields. In the episode I explain the attention, the self-attention and the multi-head attention mechanisms for both Transformers encoder and Decoder, and also the positional encoding. Go listen to this episode because it's probably my best technical episode!
Original paper: https://arxiv.org/pdf/1706.03762.pdf
Self-attention GitHub code: https://github.com/filipelauar/projects/blob/main/self_attention.ipynb
Youtube video explaining the architecture: https://www.youtube.com/watch?v=TQQlZhbC5ps
Nice blog post with code 1: http://peterbloem.nl/blog/transformers
Nice blog post with code 2: https://nlp.seas.harvard.edu/2018/04/03/attention.html
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Hey guys, in this episode I talk about the use of active learning for anomaly detection. In an anomaly detection scenario, most of the time we don't have a labeled dataset, so we have the help of an expert to annotate it. Using active learning we can hugely decrease the amount of the data to be annotated by the expert, making the job much more efficient.
Instagram: https://www.instagram.com/podcast.lifewithai/ Linkedin: https://www.linkedin.com/company/life-with-ai
Hey guys, in this episode I talk about the problem of multimodal learning, where we train a network using paired modalities of data, like image and text.
With this learning strategy we achieve more robust embeddings and enable some new applications, like the Google Multisearch.
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Hey guys, in this episode I receive as guest my professor at Sorbonne Laure Soulier. In the episode she talk a lot about Information Retrieval and NLP, which is the subject of her PhD and she current research. We discuss about conversational models (chatbots) and also explainability from Data to Text models.
In a second part of the episode, I asked her a lot of questions about her PhD and some advises that could give for people that are in their PhDs and for the ones that are undecided if they want to do it or not (me). Laure is a very good teacher and we had a really great talk, go listen to it!
Laure's linkedin: https://www.linkedin.com/in/laure-soulier-18829948/
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Hey guys, in this episode I talk about Data2Vec, a revolutionary algorithm that is able to learn from 3 different data modalities, audio, text and image and be better or comparable than previous state of the art methods. Also this algorithm uses a self-supervised approach, which means that it doesn’t use labels to the training. If you want to better understand it go listen to the episode!
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Paper: https://arxiv.org/pdf/2202.03555.pdf
Github code and models: https://github.com/pytorch/fairseq/tree/main/examples/data2vec
Hello everyone, in this episode I talk about the future of AI. I talk about applications that will impact our lives in the future, like metaverse recommendation systems, autonomous vehicles, robots, medicine and much more!
Metaverse episode: https://open.spotify.com/episode/54zU0UvAygPvmSmS8LHCmo?si=QuO39tOaSl-avJuifpf11w
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Hey guys, in this episode I have as guest Joe Scarffe, Chief Market Officer at Twine. In this episode we talk about bias an ethics in the datasets and also in the artificial intelligence algorithms. We discuss about some well known bias present in video and audio datasets like gender, age, skin color, accent and also some non trivial bias like the combination of the previous ones, the inference of the demographic attributes based on other attributes and even the existence of a qualitative bias that usually is not taken into account.
We also discuss about the ethics of using AI algorithms, where I ask Joe some really trick questions about usage of the algorithms and much more! Go listen the episode because it's a really important subject and the discussion is really interesting.
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Joe's and Twine's contact:
Twine AI Page: https://www.twine.net/ai
Twine Blog: https://www.twine.net/blog/category/ai/
Data Collector Page: https://www.twine.net/freelance/data-collector
Joe's Linkedin: https://www.linkedin.com/in/joe-s-4b0a18101/
Joe's email: joe@twine.net
Hey guys, in this episode I talk about how artificial intelligence is impacting our lives directly and indirectly, how many aspects of our life are controlled by AI algorithms. In the episode I give a lot of examples, like recommendation systems in social media, voice assistants, facial recognition, medicine and much more!
Instagram: https://www.instagram.com/podcast.lifewithai/ Linkedin: https://www.linkedin.com/company/life-with-ai
Hey guys, in this episode I talk about no code data science solution, platforms that you can create entire solutions of data science, since the data analysis and manipulation until the creation of machine learning model without writing a single line of code. In the podcast I talk about two solutions, KNIME and Dataiku, go listen to it to know their pros and cons!
Instagram: https://www.instagram.com/podcast.lifewithai/ Linkedin: https://www.linkedin.com/company/life-with-ai KNIME: https://www.knime.com/ Dataiku: https://www.dataiku.com/
Hey guys, in this episode I talk about how to make a Data Science project from scratch. In the episode I start talking about a set of steps that you need to follow to make your project, since the formulation of the hypothesis till the monitoring your model into production. After, I talk about how I applied these steps to make my fraud detection project. This episode is really practical and can help a lot of people to better formulate and develop their Data Science projects.
Instagram: https://www.instagram.com/podcast.lifewithai/ Linkedin: https://www.linkedin.com/company/life-with-ai
Hey guys, in this episode I explain all (or almost all) terms related to artificial intelligence. If you get lost in the middle of those terms that you heard and abbreviations this episode is for you, in it I explain from the basics, like what's an algorithm till the state of the art learning techniques and making connections between them. The goal of this episode is to be a dictionary to you better understand this big field of artificial intelligence.
Instagram: https://www.instagram.com/podcast.lifewithai/ Linkedin: https://www.linkedin.com/company/life-with-ai
Hey guys, in this episode I explain what the Reinforcement Learning. In the episode I give some application of where Reinforcement Learning is being used and then I explain the basic concepts of an agent, an environment, a reward, the states and transitions. Then, I talk in a really high level about the ideia behind the main algorithms like Q-Learning, Deep-Learning, Actor Critic, PPO, DDPG... In the end I talk about the advanced Reinforcement Learning techniques that we have, like curriculum learning, hierarchical learning, imitation learning, meta learning and multi agent RL.
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Github code: https://github.com/filipelauar/projects/tree/main/Reinforcement%20Learning
Hey guys, in this episode of the podcast I explain what are neural networks and how they work. In the episode I talk about the relationship between our brain and the neural network architecture and also how they are trained, using the famous algorithm of back-propagation. After understanding how neural nets work, I discuss if these "artificial intelligence" algorithm are really intelligent.
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Tensorflow playground: https://playground.tensorflow.org/#activation=tanh&batchSize=10&dataset=circle®Dataset=reg-plane&learningRate=0.03®ularizationRate=0&noise=0&networkShape=4,2&seed=0.24800&showTestData=false&discretize=false&percTrainData=50&x=true&y=true&xTimesY=false&xSquared=false&ySquared=false&cosX=false&sinX=false&cosY=false&sinY=false&collectStats=false&problem=classification&initZero=false&hideText=false
Hey guys, in this episode a have as guest the ex Oxford researcher Shrinivasan Sankar. In the episode Shrini talk about his research on Deep Learning and Active Learning, his transition from academia to an agroTech startup and also about his youtube channel, AI Bites. We also talk about how to enter/start in this Data Science and Deep Learning field.
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Shrini's Linkedin: https://www.linkedin.com/in/shrinivasan-sankar-2885016/
AI Bites YouTube channel: https://www.youtube.com/c/AIBites
Facebook's paper (Data2Vec): https://ai.facebook.com/research/data2vec-a-general-framework-for-self-supervised-learning-in-speech-vision-and-language
Hey guys, in this episode I have as guest Asya Grechka, a PhD student and my professor at Sorbonne. In the episode we mainly talk about Asya's research in generative models and GANs, but about some other interesting subjects, like general artificial intelligence. Besides being a PhD in Deep Learning, Asya also does professional climbing and was 2x finalist of the American Ninja Warrior, in the episode she told us about the experience.
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Github code about image editing made by Asya:
Hey guys, in this episode I talk about the pre training process of the state of the art NLP and computer vision transformer architectures. Since 2017 we train NLP (BERT, GPT, ELECTRA) networks with a masked language model using a self-supervised procedure, and now (since 2022) we are also able to train vision (MAE) networks using the same masked language model procedure. This way of self-supervised pre training enable us to train accurate models that really understands semantic and context without labeled data. I also talk about a tabular transformed architecture (TabTransformer - 2020) using the same approach achieve state of the art results compared to ensemble methods.
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
BERT paper: https://arxiv.org/pdf/1810.04805.pdf
GPT3 paper: https://arxiv.org/pdf/2005.14165.pdf
ELECTRA paper: https://arxiv.org/pdf/2003.10555.pdf
MAE paper: https://arxiv.org/pdf/2111.06377.pdf
TabTransformers paper: https://arxiv.org/pdf/2012.06678.pdf
Hey guys, in this episode I talk about the how to attack a neural network and also how to defend your network against these attacks. In the episode I give the example of attacking autonomous vehicles, how people manage to create images that make this vehicle think that a person is a tree or a car, for instance, explaining also the methods that do it. I also talk about how to defend your network from these attacks using different techniques of adversarial defenses.
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Code: https://github.com/filipelauar/projects/blob/main/adversarial_attacks_pytorch.ipynb
Hey guys, in this episode I explain how an autonomous vehicle can learn to drive in a specific city and be able to generalize the driving knowledge to any other city. To explain it, I show the concept of of semi-supervised learning and domain adaptation. I explain the the ideia with the first proposed architecture and what we have today as state of the art for this problem.
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Code: https://github.com/filipelauar/projects/blob/main/domain_adaptation_semi_supervised_learning_MNIST_pytorch.ipynb
Hey Guys, in this episode I have as guest my fiend Miguel to talk about the problem of speech to text. The problem is basically transcript an audio to text, in the episode we talk about many application, a little bit of the history of the problem and also about the state of the art proposed by Facebook that uses self-supervised learning.
Instagram: https://www.instagram.com/podcast.lifewithai/ Linkedin: https://www.linkedin.com/company/life-with-ai Linkedin do Miguel: https://www.linkedin.com/in/lucasagrizzi/
Hey guys, in this episode of the podcast I have as guest my friend Vinicius Bonemer to talk about the virtual world a.k.a metaverse. In the episode we go deep into the ideia/concept of a virtual world/metaverse and explain the different between Facebook's Metaverse and also NVIDIA's Omniverse. We discuss a lot of possibilities about this new universe, what we already have, opportunities, challenges, algorithms, interoperability, NFTs and much more...
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Vinicius's likedin: https://www.linkedin.com/in/vbonemer/
NVIDIA's Omniverse keynote: https://www.youtube.com/watch?v=RPfItXmIRG4
Facebook's Metaverse keynote: https://www.youtube.com/watch?v=gElfIo6uw4g
Wall Street Journal podcast about Facebook: https://open.spotify.com/playlist/0Amco5yINtnBGIY94Z2aHb?si=Yn9NoiVPSFSca3UOFLx4Iw
Wall Street Journal article about Facebook: https://www.wsj.com/articles/the-facebook-files-11631713039
Hey guys, in this episode I receive as guest my friend Vinicius Akira to talk about our experience to find our internships here in France and we also give a lot of tips to help you get yours. We talk about the different parts of the selective process, like how to prepare and send your CV, how to pose yourself for the HR interview and how to prepare to the technical interviews.
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Akira's Linkedin: https://www.linkedin.com/in/daquisu/
Hello everyone, in this episode I interviewed Matthieu Cord, my professor at Sorbonne University. Matthieu is a researcher at Sorbonne in the field of Deep Learning and Computer Vision and also a researcher at Valeo's AI lab, a french autonomous driving company. In the episode we talk mainly about Matthieu's current research topics, autonomous driving and the hottest topics in both research and industry.
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Matthieu's website: http://webia.lip6.fr/~cord/
Hey guys, in this episode I receive as a my friend Franco and we talk about energy efficiency, image enhancement and active learning. Franco told about his findings working on a project an energy efficiency project for the Uruguay government, image enhancement with reinforcement learning at Huawei and now his PhD on the active learning field.
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Franco's Linkedin: https://www.linkedin.com/in/franco-marchesoni/
Hey guys, in this episode I talk about how is being the experience of have the podcast. I talk about my routine as a podcaster, the equipment and tools that I use to record and edit and also how the podcast is helping me to improve my hard and soft skills.
Instagram: https://www.instagram.com/podcast.lifewithai/ Linkedin: https://www.linkedin.com/company/life-with-ai
Hey guys, in this episode I talk about the main points of the episode that I recorded in Portuguese with Fernando Santos, a post doc researcher and professor at USP. In this episode, I talk about tips to train neural networks and some techniques that are really improving the convergence of the training, like data augmentation and self-supervised learning.
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Fernando's Linkedin: https://www.linkedin.com/in/fernando-persan/
Tutorial's code: https://github.com/maponti/trainingdeepnetworks
Hey guys, in this episode I talk about how to input tabular data into artificial intelligence algorithms. You can understand tabular data as the famous excels tables, where you have categorical and numerical variables. in this episode I explain how the algorithms see tabular data and have how to transform this into the right format.
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Code: https://github.com/filipelauar/projects/blob/main/categorical_encoding_numerical_scaling_normalization.ipynb
Hey guys, in this episódio I have as guest my friend Breno Skuk to talk about binary neural networks. Binary Neural Networks can decrease the inference time of the model by 40x and the memory size by until 64x! In this episode we talk about many pros and cons of the use of these networks and how its evolution and implementation can impact our lives.
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Code: https://github.com/filipelauar/projects/blob/main/binary_neural_networks.ipynb
Breno's Linkedin: https://www.linkedin.com/in/breno-baldas-skuk-117213140/
Hey guys, in this episode I explain the ideia behind 3 very important algorithms widely used in Natural Language Processing (NLP) and time series. I explain classic RNN, GRU and LSTM algorithms, explaining the ideia behind each one of them and also the goal of the famous and very important LSTM and GRU gates.
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Code: https://github.com/filipelauar/projects/tree/main/RNN_applications
Hey guys, in this episode I talk about applications related to RNNs in the real world. I talk about applications of Natural Language Processing (NLP), like text classification, chatbots, text generation, speech recognition... and also applications related to time series, like weather forecasting, predictive maintenance and much more!
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Code: https://github.com/filipelauar/projects/tree/main/RNN_applications
Hey guys, in this episode I talk about applications of GANs, Generative Adversarial Networks, in the real world. I talk about many applications in photo and video editing, super resolution, video games, autonomous driving and much more!
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Code: https://github.com/filipelauar/projects/tree/main/GAN_applications
In this episode I talk about different applications of artificial intelligence in the real world using CNN networks. I talk about 3 main problems, classification, detection and segmentation, explain what's being done in each of these problematic and also the state of the art to solve them.
Website to find state of the art models: https://paperswithcode.com/sota/
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Code: https://github.com/filipelauar/projects/tree/main/CNN_applications
In this episode I'm with my friend Gabriel Oliva to talk about the use o Artificial Intelligence in the financial market. We talk about many applications, such as satellite analysis, price prediction, assets allocation and many more. We also discuss the use of AI more as a tool than a final solution.
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Gabriel's Linkedin: https://www.linkedin.com/in/gabriel-oliva-199359148/
In this episode I'm with my friend Kevin Kühl to talk about the first steps to get into the financial market. We talk about different types of investments like fixed incomes and variable incomes, the stocks, the REITS and also the ETFs. If you want to now more about the financial market and is thinking to get into it, this episode is for you!
https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Kevin's Linkedin: https://www.linkedin.com/in/kevinfkuhl/
In this episode I have as a guest my friend Kevin Kühl, which worked with modeling complex systems in the ICMC lab from USP in partnership with a lab at the Imperial College London. In the episode we talk about how modeling complex systems can help us understand the behavior of the human brain, the isolation system against Covid19 and even the hands clapping in a birthday party. Instagram:
https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Kevin's Linkedin: https://www.linkedin.com/in/kevinfkuhl/
In this episode I talk about the problems of dimensionality reduction and clustering. I explain the applications of each one of these problems and also the most famous methods for solving them, such as PCA, KPCA, ICA and NNMF for the dimensionality reduction and the Kmeans for the clustering problems. In the end I also explain the autoencoders, which are powerfull neural networks that can be used for both problems.
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Code: https://github.com/filipelauar/projects/tree/main/dimensionality%20reduction%20and%20clustering
In this episode I have as a guest my friend Vinícius Bonemer, that worked with augmented reality and virtual reality in the Apple lab in Brazil. In this conversation we explain what is AR and VR, the differences between them, many applications and also talk about the future of the fields!
In this episode I explain a new class of problems that are able to be solved from the neural network CycleGAN. With this new network, we are able to transform an image into another, such as transform a horse into a zebra, a cat into a dog of even a normal picture into a famous painting made by Van Gogh or Monet!
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Code: https://github.com/filipelauar/projects
In this episode I explain the solution I had to win a kaggle competition in my university (Telecom Paris). I explain the different solutions I had and what I did to improve the results. In the end I also talk about a possible solution that could win against my own using ensembles.
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
GitHub with the solution: https://github.com/filipelauar/Kaggle/blob/main/cancer_detection_challange_notebook.ipynb
In this episode I give various examples about racist algorithms in different domains like face recognition, healthcare insurance, crime probability, hate speech and much more! I talk not only about racial problems, but also about gender and religion problems caused by artificial intelligence algorithms.
In this episode I explain how the machine learning algorithms based on trees and forests work. In the episode I explain in a simple way the idea behind the decision trees and also the powerful algorithms RandomForest and XGBoost and its respectives methods, bagging and boosting.
In this episode we talk about some myths and truths about Artificial Intelligence. We talk about the general artificial intelligence, where the robots would conquer the earth, and a lot of other curiosities about AI!
In this episode I talk about the next generation of instagram filters, that we will be able to edit easily specific attributes of our face or even mix two people. I also explain the artificial intelligence networks behind this next generation of filters.
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
Code: https://github.com/filipelauar/projects/tree/main/styleGAN_interfaceGAN_IDInvert_faceAttributeFAN
In this episode I talk about how artificial intelligence is generating people that doesn't exist. I explain how the GANs work, the Generative Adversarial Networks that are able to generate perfectly people from noise.
Instagram: https://www.instagram.com/podcast.lifewithai/
Linkedin: https://www.linkedin.com/company/life-with-ai
This person is real: https://thispersondoesnotexist.com/
Which face is real: https://www.whichfaceisreal.com/
In this episode I talk about how Facebook algorithms work. I explain the main goal of the platform and also how do they use our data to make money. I also give some examples of how powerful data is.
In this episode we explain how computers understand texts. I talk about a representation based on the dictionary and a more efficient one based on the attributes/characteristics of the word so that the computer can understand them better.
In this episode I talk about artificial intelligence for image. I explain the idea behind one of the most popular AI algorithm, the neural networks and show how it's able to learn.
In this episode of the podcast Life with AI I talk about the problem of predicting news. In this episode I show how I statistically proved that it's possible to predict news and I also show could we use such algorithm in practice.
This is the first episode of the podcast Life with AI, the podcast that talks about artificial intelligence in a simpler way. In this episode I explain the problem of detecting insider trading frauds and how I did to create a dataset of numbers of possible insider trading frauds for the Brazilian market. Hope you like it!