El poscats ejes del pensamiento lógico matemático en un mundo digital para la educación infantil a los maestros y futuros maestros a desarrollar en los niños y niñas habilidades lógico matemáticas desde educación inicial, ya que estas son fundamentales para toda su vida, permitiendo a los niños resolver problemas, tomar decisiones y entender el mundo que les rodea.

Just hours after Donald Trump claimed victory in the US presidential election, rumours started swirling that something was afoot. A graph went viral on social media that appeared to show there were 20 million more votes cast in 2020 than in the 2024 election. Where had these supposedly “missing” votes gone? Conspiracy theorists on both sides of the political spectrum began shouting claims of fraud. The answer, it turns out, is rather more straightforward. Presenter: Charlotte

⚖️ Scaling Laws for PrecisionThis research paper investigates the impact of precision in training and inference on the performance of large language models. The authors explore how precision affects the effective parameter count and propose scaling laws that predict performance degradation due to low-precision training and post-training quantization. They find that overtrained models are more sensitive to post-training quantization, and that training larger models in lower precision might be

In this Tier 1 Interventions podcast episode, Cheri shares the concept of "flow" as a transformative approach to classroom learning. By unpacking the neuroscience of brain waves, the importance of sleep, and responsive teaching methods, she and Jonily explore how educators can adapt their strategies to create engaging and effective lessons. The discussion also highlights practical techniques like brain interruption activities and strategies for embedding responsive teaching into your daily

Levels of AI Information Processing and Pattern RecognitionThe sources explore how the varying capabilities and limitations of AI systems reflect different levels of information processing and pattern recognition.Basic Information Processing: Some AI, such as Gemini, initially exhibit linear processing and standard scientific skepticism. These systems struggle with pattern recognition, missing connections to recent research and empirical evidence. However, with sufficient context, they can

Are academically challenging goals appropriate for students with dyscalculia? YES! Find out why, and how to implement them, in this episode.

⌛️ The Surprising Effectiveness of Test-Time Training for Abstract ReasoningThis paper examines how test-time training (TTT) can enhance the abstract reasoning abilities of large language models (LLMs). TTT, which updates model parameters during inference, significantly improves performance on the Abstraction and Reasoning Corpus (ARC) benchmark. Key factors for effective TTT include initial fine-tuning, auxiliary tasks, and instance-specific training. The approach achieves state-of-the-art results

A huge quantity of clothing is produced every year around the world. But is so much made that there are already enough tops, trousers, skirts and all the rest to clothe humanity for decades into the future? That’s a claim that has been percolating around the internet recently, that there are already enough clothes for the next six generations. Tim Harford and Beth Ashmead Latham explore the source of this claim and, with help from Sabina Lawreniuk from Nottingham University, find that the evidence

Dr. Wall shares exciting, easy, and impactful IEP or 504 goals in this episode. She shares information from David Bateman and Dr. Lisa Goran, delivered at the 2024 CASE conference. Don't miss this one!

Podcast sobre simplificación de fracciones

Great breakdown on the AI

Shows a way to add constraints to an optimization problem which wishes to bound the sum of the 'k' largest components of a given collection whiles minimizing a given objective.

Examines the problem AX = b + epsilon looking for an unbiased estimator. when using "trace" norms.Shows that the estimator is always the same.The infinite collection of these results imply that the usual "best" estimator has the property that its Maximum Eigenvalue is the smallest amongst all unbiased linear estimators.

It is known how to compute the probability distribution of a variable that is transformed in a non-singular way.When this is not the case, people find ad hoc ways of computing the probability distribution for singular transformations.The paper shows a more systematic process that leads to formulas form complex singular transformations.

Shows how balance laws over arbitrary regions leads to a "local" version which is a Partial Differential Equation.

Looks at measure theoretic conditional expectation in a discrete setting using the familiar example of single 6 sided dice.Shows how measure theory helps with the notion of conditional expectation when the conditional event has probability 0.

Tries to motivate Lebesgue Integration by trying to find an improved limit theorem for Riemann Integration.In the process, find that this seems a lot like trying to find limit results when only dealing with rational numbers.End up with a example where we can "see" what the integral is, but Riemann does not have an answer.From the example see that by describing the unusual function by its range rather than its domain, we can compute the integralProblem is that, in general we need to find the length

Defines fractal dimensionWhy it is a useful concept.Computes the fractal length of a simple non-fractal setComputes the "length" of a self-similar fractal in its "natural" fractional dimension.

Inner Product can be thought of coming from the notion of Projection.If follows from doing a Mathematical "re-factoring", something that one usually only hears in the context of software engineering.

Rather than giving a formula for the determinant, one tries to derive it from first principles.The principles come about by looking at the volume of a parallelepiped.Uses the properties that such a formula would have to derive it.