Associate Prof. (Prof. Titular de Universidad). Coordinator Bachelor in Data Science and Engineering @URJC. Data engineering @DSLAB_URJC
Madrid, Spain
Joined May 2008
- Tweets2.9K
- Following719
- Followers936
- Likes2.3K
Felipe Ortega retweeted
I’m very excited to share @nvidia Kumo Tabular, a new family of foundation models for tabular data.
Kumo Tabular establishes the new Pareto frontier across the entire accuracy–inference-time tradeoff.
Just as importantly, we are releasing it openly: open weights, open-source software, and a permissive license for commercial use.
HuggingFace: huggingface.co/nvidia/Kumo-T…
GitHub: github.com/NVIDIA/structured…
Felipe Ortega retweeted
I put together a mega write-up on GPT-6 Astra & looped transformers.
How looped transformers / recurrent depth works, cost-tradeoffs, whether it hides reasoning traces, with lots of figures and a tour of recent looped transformer research.
Felipe Ortega retweeted
Introducing the GeoLibre R package!
GeoLibre already runs as a desktop app, a web app, and a Jupyter widget. Now it runs in R too. The new geolibre R package is an htmlwidgets interface that embeds the full GeoLibre GIS in the RStudio Viewer, Quarto, R Markdown, and Shiny.
What you get:
- A real GIS, not a static map. Layer panel, styling, drawing tools, identify, and the processing toolbox are all there inside your R session.
- Bring your own data. `sf` objects, GeoJSON, CSV and XY tables, Shapefile, GeoPackage, FlatGeobuf, GeoParquet, KML, Cloud Optimized GeoTIFF, XYZ, WMS, WMTS, WFS, PMTiles, vector tiles, 3D Tiles, and even video overlays.
- Shiny, in both directions. `geolibre_proxy()` updates a live map without rebuilding it, and drawn features, selected features, identify results, and the current view come back as Shiny inputs. You can even run processing algorithms from the server.
The package is on GitHub now while the first CRAN submission is prepared.
Try it out
- R package website: r.geolibre.app
- Live interactive example: r.geolibre.app/articles/inte…
- Source: github.com/opengeos/geolibre…
- GeoLibre Web: web.geolibre.app
- Documentation: geolibre.app
#RStats #GIS #Geospatial #OpenSource #Shiny #MapLibre #GeoLibre
Felipe Ortega retweeted
Just saw that the LLMs-from-scratch repository passed 100,000 stars on GitHub!
This is super cool and motivating. I am really happy to see that this open-source repo has helped so many people.
Thanks also to everyone who shared ideas and opened PRs with improvements!
Of course, I plan to keep adding new material, including new attention variants and architectures (while bigger projects like RL and Reasoning From Scratch live in their separate repositories).
I am also currently working on a larger applied custom “small” LLM project. It has been keeping me super busy this month, but I will share more on that soon.
If you are new to it, some of the highlights include
1. Of course, the complete code path from tokenization and attention to pretraining, classification, and instruction fine-tuning, etc. All of it FROM SCRATCH, of course! (RL lives in a companion repo.)
2. From-scratch implementations of Llama, Qwen, Gemma, and Olmo (smaller variants that run locally and can be plugged into the training scripts).
3. From-scratch implementations of attention alternatives and other architecture components, such as GQA, MLA, sliding-window attention, Gated DeltaNet, DeepSeek Sparse Attention, cross-layer KV sharing, and mixture-of-experts
4. Materials on KV caching, training performance, memory-efficient weight loading, DPO, evaluation, and LoRA
So, if you don’t have any weekend plans yet, happy tinkering!
Felipe Ortega retweeted
After 18 months of writing, coding, and experimenting, Build a Reasoning Model (From Scratch) is
finally out!
My first copies just arrived! 📚
440 full-color pages. Inference scaling, reinforcement learning, and distillation from scratch.
Felipe Ortega retweeted
BIG claim from new MIT + Oxford + Carnegie Mellon and other top labs paper:
AI can boost performance at first and then leave people less able to think through problems on their own.
Just minutes of AI help can improve scores now while weakening independent problem-solving right after.
The interesting part is that the damage is not just lower accuracy.
It is lower persistence, which is usually the hidden engine of learning, because skill grows through repeated contact with difficulty, not just exposure to correct answers.
That's why a good teacher sometimes withholds help to preserve struggle as part of the lesson, while today’s chatbots are tuned to erase friction on demand.
Across 3 experiments in math and reading, about 1.2K people either worked alone or used a GPT-5-based assistant for part of the task.
Assisted users finished early questions faster, but after roughly 10 minutes without AI, they solved less, stalled more, and quit sooner.
That happens because hard thinking is not only about getting answers; it is also about building the habit of holding a problem in mind, testing steps, and pushing through confusion.
The sharpest drop came from people who used the model for direct answers, not from those who used it more like a hint system, which suggests the real issue is not AI exposure itself but replacing effort with completion.
The result is not that AI makes people less capable by default, but that answer outsourcing can shrink the mental effort that normally trains skill.
----
Paper Link – arxiv. org/abs/2604.04721
Paper Title: "AI Assistance Reduces Persistence and Hurts Independent Performance"
Felipe Ortega retweeted
Mientras veíamos a compañeros hacer muchas cosas en Claude nosotros hemos estado trabajando en una línea principal, la metodología. Y el/los métodos radican en las skill. Y una vez más en mi última reunión hablé de los incentivos "perversos" que tanto gusta a @Recuenco
Y es que se avecinan guerras civiles en las grandes corporaciones, una skill es desgranar lo máximo posible años y años de trayectoria y conocimiento, y si pedir una query ya generaba fricción, imaginad lo que está suponiendo pensar en compartir skills. Empresas con EREs cada 2-3 años, cambios, etc...Nosotros hemos apostado por construir y compartir, sabiendo que cometemos un riesgo enorme de fuga de conocimiento, pero no hemos venido a meternos en una trinchera, si yo quiero crecer necesito que mis compañeros crezcan, y así la empresa.
Pero bueno, ya oigo cosas referidas a hashear campos de tablas para que cueste más encontrar información, información encriptada entre compañeros, etc.
Dois engenheiros da Anthropic acabaram de mudar a forma como devs pensam sobre IA.
Barry Zhang e Mahesh Murag subiram no palco do AI Engineer Code Summit e disseram uma frase que incomodou muita gente:
"Parem de construir agentes. Construam Skills."
Em 16 minutos eles provam que a indústria inteira está resolvendo o problema errado.
Aqui está o que a maioria não entendeu:
→ Skills são pastas. Literalmente pastas com arquivos markdown.
→ Elas ensinam ao Claude o SEU fluxo de trabalho, a SUA expertise, o SEU domínio.
→ Um único agente genérico + biblioteca de Skills específicas supera dezenas de agentes especializados.
→ Fortune 100s já estão deployando Skills em escala pra ensinar agentes sobre processos internos.
→ Times de produtividade com 10.000+ devs usam Skills pra padronizar como código é escrito.
A analogia que eles usaram é perfeita:
Quem você quer fazendo seu imposto de renda? O gênio com QI 300 que nunca viu legislação tributária, ou o contador experiente que faz isso há 20 anos?
Inteligência sem expertise é entretenimento.
Expertise empacotada é produtividade.
O que mudou: a Anthropic parou de tentar criar agentes diferentes pra cada domínio.
Perceberam que com Claude Code, o padrão é sempre o mesmo. Um modelo acoplado a um runtime com filesystem.
A diferença entre um agente medíocre e um extraordinário não é o modelo. É o conhecimento de domínio que você alimenta.
Skills resolvem isso com progressive disclosure. O agente só carrega o nome e descrição da skill. Quando relevante, puxa o SKILL.md. Quando precisa de mais, navega os arquivos de referência. Zero desperdício de contexto.
Isso não é uma feature. É uma mudança de paradigma.
Quem entender isso agora vai operar em outro nível daqui a 90 dias.
Quem ignorar vai continuar escrevendo prompts de mil palavras toda vez que abrir o chat. E ainda explicar de novo e de novo o que “realmente” quer.
Gracias a la promo de @JMachicado_ he descubierto @ElTriplazoTV y el episodio con @Oscar_GarriLedo. No soy objetivo en este tema, porque sé de la importancia de la #CienciadeDatos en la toma de decisiones en deporte 🏀 🏈⚽️ Os seguiré con interés.
Por cierto (cuña publi): también formamos a este tipo de perfiles: blogs.etsii.urjc.es/musa/ Por ahora, más enfocado a ⚽️, pero los amantes del 🏀 también estamos por ahí y hay muchas aportaciones interesantes por hacer
@GarciaAller Para alguna #Pausa futura: cuando se sugiera que España tiene que potenciar el consumo de electricidad, porque tenemos mucha producción de renovables, etc. Mapa: foroindustriayenergia.com/ma… El problema (uno de ellos) es que la red de distribución está saturada
Felipe Ortega retweeted
An Open Letter to Georgetown Students, In Response to Recent Announcements by the University about “Generative AI”
medium.com/center-on-privacy…
For all people willing to really *understand* (of many kinds) models in data science, including recent advances in interpretation and evaluation, this is a key resource: m-clark.github.io/book-of-mo… Big pro: all examples in both R and Python, because focus is on mastering concepts.
Another wonderful title from @CRC_MathStats Data Science Book Series. I have read Michael Clark's blog for many years, and it is also full of (expanded) examples and insights on data modelling: m-clark.github.io/
Felipe Ortega retweeted
MIT is offering Books on AI & ML (ABSOLUTELY FREE):
1. Foundations of Machine Learning
cs.nyu.edu/~mohri/mlbook/
2. Understanding Deep Learning
udlbook.github.io/udlbook/
3. Introduction to Machine Learning Systems
mlsysbook.ai/assets/download…
4. Algorithms for ML
algorithmsbook.com
5. Deep Learning
deeplearningbook.org
6. Reinforcement Learning
andrew.cmu.edu/course/10-703…
7. Distributional Reinforcement Learning
direct.mit.edu/books/oa-mono…
8. Multi Agent Reinforcement Learning
marl-book.com
9. Agents in the Long Game of AI
direct.mit.edu/books/oa-mono…
10. Fairness and Machine Learning
fairmlbook.org
11. Probabilistic Machine Learning
❯ Part 1 : probml.github.io/pml-book/bo…
❯ Part 2 : probml.github.io/pml-book/bo…
Felipe Ortega retweeted
If you are looking for reading material for the upcoming weekend, (the long-promised) Chapter 5 on LLM self-refinement is now finally out in the early access.
Here, we continue the inference-time scaling theme, but we move beyond self-consistency and voting. More specifically, we implement a self-refinement loop, where a model iteratively critiques and improves its own answers.
Along the way, the chapter implements several core pieces that will become essential in the upcoming reinforcement learning chapters like log-probability scoring (I finished the RL chapter last week, and it will hopefully be out soon, too).
Like always, all of this is implemented from scratch, step by step, of course. I think that seeing it all in working code really helps with understanding how LLM reasoning methods work (versus just looking at the equations).
Anyways, with Chapter 5 out, the early-access version of the book has grown quite a bit (~300 pages) and now finally wraps up inference-time reasoning before we transition into the (even more) fun part: reinforcement learning (from scratch!) in the next chapter.
🔗 Here's the link to the book's early access: mng.bz/dWGv
Happy reading
(PS: Sorry for the delay. It turns out the submitted manuscript got a bit backlogged in the publisher’s processing pipeline. The silver lining is that the reinforcement learning chapter hopefully now follows sooner! )
Jensen Huang: “The world needs to build energy infrastructure as fast as it builds AI infrastructure”. Why don’t we start investing more effort in identifying and using energy-efficient algorithms? Innovation is in reducing energy consumption not building more power plants #AI
Se veía venir capacityglobal.com/news/ai-b…
Felipe Ortega retweeted
¡Manic miner en modo texto! En esta prueba se mueve el personaje en un escenario inspirado en la pantalla 1 del juego
Lo veo interesante para enseñar programación con python de forma divertida, sin tener que utilizar bibliotecas gráficas
github.com/Obijuan/unicode_p…
Felipe Ortega retweeted
The 3rd edition of my book Deep Learning with Python is being printed right now, and will be in bookstores within 2 weeks. You can order it now from Amazon or from Manning.
This time, we're also releasing the whole thing as a 100% free website.
I don't care if it reduces book sales, I think it's the best deep learning intro around, and more people should be able to read it.
