boardsofdata's profile picture. science & startup worker. AI, biotech, manufacturing. @stammbio @udesa

martin palazzo

@boardsofdata

science & startup worker. AI, biotech, manufacturing. @stammbio @udesa

Fijado

Molecular biology 🧬 wet lab iteration for cell therapies implies intense R&D resources. By leveraging AI, computational biology algorithms and multi-omics assay data @StammBio presents MoNA: a cell representation atlas designed to accelerate bio-innovation cycles. Take a look👇

1/ 🚨 Uncertainty in cell & gene therapy development? Enter MoNA: our Multi-Omic Network Atlas technology!

StammBio's tweet image. 1/ 🚨 Uncertainty in cell & gene therapy development? Enter MoNA: our Multi-Omic Network Atlas technology!


martin palazzo reposteó

Data often lie on a low-dimensional manifold embedded in a high-dimensional space. But these manifolds are often highly non-linear, making linear dimensionality reduction methods like PCA insufficient. This has motivated the development of non-linear dimensionality reduction.


Is this a circular RNA reference ? 😅

wait for it



martin palazzo reposteó

AI ppl: we basically solved bio with the thinking sands Bio ppl: after decades of lab toil, we solved how to stably express GFP in stem cells

1/ Happy to share important work done with my co-author Andrew Khalil in the labs of Rudolf Jaenisch @WhiteheadInst @MITBiology @MIT and Dave Mooney @Harvard @wyssinstitute trying to assess and fix the major problem of transgene silencing in human ESC/iPSC based work



martin palazzo reposteó

We’re moving biology from the wet lab to the gpu cluster.


martin palazzo reposteó

I find it fascinating that momentum in standard convex optimization is just about making convergence faster, but in nonconvex problems, it's sometimes the only way a method can work at all. Just saw a new example of this phenomenon in the case of difference-of-convex functions.

konstmish's tweet image. I find it fascinating that momentum in standard convex optimization is just about making convergence faster, but in nonconvex problems, it's sometimes the only way a method can work at all. Just saw a new example of this phenomenon in the case of difference-of-convex functions.

martin palazzo reposteó

Graphons are mathematical objects that model the structure of massive networks. In machine learning, they provide a powerful framework for analyzing and generating large graphs. They are used to estimate the underlying structure of a network, predict missing links, and understand…

probnstat's tweet image. Graphons are mathematical objects that model the structure of massive networks. In machine learning, they provide a powerful framework for analyzing and generating large graphs. They are used to estimate the underlying structure of a network, predict missing links, and understand…

martin palazzo reposteó

Theorem: The maximum possible duration of the computational singularity is 470 years. Proof: The FLOPs capacity of all computers which existed in the year 1986 is estimated to be at most 4.5e14 (Hilbert et al. 2011). Based on public Nvidia revenue and GPU specs, this capacity…


martin palazzo reposteó

How does neural feature geometry evolve during training? Modeling feature spaces as geometric graphs, we show that nonlinear activations drive transformations resembling discrete Ricci flow - revealing how class structure emerges and suggesting geometry-informed training…

mweber_PU's tweet image. How does neural feature geometry evolve during training? Modeling feature spaces as geometric graphs, we show that nonlinear activations drive transformations resembling discrete Ricci flow - revealing how class structure emerges and suggesting geometry-informed training…

martin palazzo reposteó

The simplex method is an algorithm that turns an optimization problem, like setting up an investment portfolio, into a geometry problem. Recently, the scientists Sophie Huiberts (left) and Eleon Bach reduced the runtime of the simplex method. quantamagazine.org/researchers-di…

QuantaMagazine's tweet image. The simplex method is an algorithm that turns an optimization problem, like setting up an investment portfolio, into a geometry problem. Recently, the scientists Sophie Huiberts (left) and Eleon Bach reduced the runtime of the simplex method. quantamagazine.org/researchers-di…
QuantaMagazine's tweet image. The simplex method is an algorithm that turns an optimization problem, like setting up an investment portfolio, into a geometry problem. Recently, the scientists Sophie Huiberts (left) and Eleon Bach reduced the runtime of the simplex method. quantamagazine.org/researchers-di…

martin palazzo reposteó

Do more math.


martin palazzo reposteó

🚨 We wrote a new AI textbook "Learning Deep Representations of Data Distributions"! TL;DR: We develop principles for representation learning in large scale deep neural networks, show that they underpin existing methods, and build new principled methods.

druv_pai's tweet image. 🚨 We wrote a new AI textbook "Learning Deep Representations of Data Distributions"!   

TL;DR: We develop principles for representation learning in large scale deep neural networks, show that they underpin existing methods, and build new principled methods.

martin palazzo reposteó

Two sources, two frequencies #mathart


martin palazzo reposteó

A 7 million parameter model from Samsung just outperformed DeepSeek-R1, Gemini 2.5 Pro, and o3-mini on reasoning benchmarks like ARC-AGI. Let that sink in. It’s 10,000x smaller yet smarter. The secret is recursion. Instead of brute-forcing answers like giant LLMs, it drafts a…

VraserX's tweet image. A 7 million parameter model from Samsung just outperformed DeepSeek-R1, Gemini 2.5 Pro, and o3-mini on reasoning benchmarks like ARC-AGI.

Let that sink in.
It’s 10,000x smaller yet smarter.

The secret is recursion.
Instead of brute-forcing answers like giant LLMs, it drafts a…

martin palazzo reposteó

*The Origins of Representation Manifolds in LLMs* by Modell et al. They study the presence of "interpretable features" in LLMs embedded as manifolds and how their geometry connects to the internal representations of the models. arxiv.org/abs/2505.18235

s_scardapane's tweet image. *The Origins of Representation Manifolds in LLMs*
by Modell et al.

They study the presence of "interpretable features" in LLMs embedded as manifolds and how their geometry connects to the internal representations of the models.

arxiv.org/abs/2505.18235

martin palazzo reposteó

This new research showed empirically that KL divergence is the most accurate predictor of catastrophic forgetting in LLMs! And since forgetting tracks KL drift from the base, methods that keep KL small tend to forget less too.

askalphaxiv's tweet image. This new research showed empirically that KL divergence is the most accurate predictor of catastrophic forgetting in LLMs!

And since forgetting tracks KL drift from the base, methods that keep KL small tend to forget less too.

martin palazzo reposteó

HOW INFORMATION FLOWS THROUGH TRANSFORMERS Because I've looked at those "transformers explained" pages and they really suck at explaining. There are two distinct information highways in the transformer architecture: - The residual stream (black arrows): Flows vertically through…

repligate's tweet image. HOW INFORMATION FLOWS THROUGH TRANSFORMERS
Because I've looked at those "transformers explained" pages and they really suck at explaining.

There are two distinct information highways in the transformer architecture: 
- The residual stream (black arrows): Flows vertically through…
repligate's tweet image. HOW INFORMATION FLOWS THROUGH TRANSFORMERS
Because I've looked at those "transformers explained" pages and they really suck at explaining.

There are two distinct information highways in the transformer architecture: 
- The residual stream (black arrows): Flows vertically through…
repligate's tweet image. HOW INFORMATION FLOWS THROUGH TRANSFORMERS
Because I've looked at those "transformers explained" pages and they really suck at explaining.

There are two distinct information highways in the transformer architecture: 
- The residual stream (black arrows): Flows vertically through…

KV caching overcomes statelessness in a very meaningful sense and provides a very nice mechanism for introspection (specifically of computations at earlier token positions) the Value representations can encode information from residual streams of past positions without…



martin palazzo reposteó

💡 Me encantaron las sesiones de pósters en el Simposio Científico de IA y Aplicaciones #SCIAA2025! Había muchos trabajos interesantes. En particular quiero destacar dos proyectos que me llamaron la atención: 🔹 Impacto del Prompt Engineering en el rendimiento de LLMs: estudio…

ch4rleston's tweet image. 💡 Me encantaron las sesiones de pósters en el Simposio Científico de IA y Aplicaciones #SCIAA2025!

Había muchos trabajos interesantes. En particular quiero destacar dos proyectos que me llamaron la atención:

🔹 Impacto del Prompt Engineering en el rendimiento de LLMs: estudio…
ch4rleston's tweet image. 💡 Me encantaron las sesiones de pósters en el Simposio Científico de IA y Aplicaciones #SCIAA2025!

Había muchos trabajos interesantes. En particular quiero destacar dos proyectos que me llamaron la atención:

🔹 Impacto del Prompt Engineering en el rendimiento de LLMs: estudio…

Que capo

At 26, during the Reign of Terror in France, Jean-Baptiste Joseph Fourier narrowly avoided the guillotine. A decade later, he made a discovery that changed mathematics forever. @shalmawegs reports: quantamagazine.org/what-is-the-fo…



martin palazzo reposteó

The most complex biological system that we meaningfully understand is a single virion. Everything more complex than that (including bacteria, single human cells, animals, and of course, humans) is totally beyond our comprehension.


Llego el dia de decidir entre: - Bosch - Makita - DeWalt - Black & Decker


Loading...

Something went wrong.


Something went wrong.