MiniJinja is a powerful but minimal dependency template engine for Rust which is based on the syntax and behavior of the Jinja2 template engine for Python.

It's implemented on top of serde and only has it as a single dependency. It supports a range of features from Jinja2 including inheritance, filters and more. The goal is that it should be possible to use some templates in Rust programs without the fear of pulling in complex dependencies for a small problem. Additionally it tries not to re-invent something but stay in line with prior art to leverage an already existing ecosystem of editor integrations.

#rust #jinja2 #templates

Researchers discover a material that can learn like the brain

Vanadium Dioxide (VO2), a compound used in electronics, is capable of “remembering” the entire history of previous external stimuli. This is the first material to be identified as possessing this property, although there could be others.
In his experiments, Samizadeh Nikoo applied an electric current to a sample of VO2. “The current moved across the material, following a path until it exited on the other side,” he explains. As the current heated up the sample, it caused the VO2 to change state. And once the current had passed, the material returned to its initial state. Samizadeh Nikoo then applied a second current pulse to the material, and saw that the time it took to change state was directly linked to the history of the material. “The VO2 seemed to ‘remember’ the first phase transition and anticipate the next,” explains Prof. Elison Matioli, who heads the POWERlab. “We didn’t expect to see this kind of memory effect, and it has nothing to do with electronic states but rather with the physical structure of the material. It’s a novel discovery: no other material behaves in this way.”

Well...
That's how Gemini AI works - it's Google's newest LLM.

With all due respect to ChatGPT and Gigachat….

A little scary, to be honest... 😅

@science - share the post please 👌

A Chinese startup, Betavolt, has introduced an innovative 'atomic energy battery' that can generate electricity for 50 years without needing recharging or maintenance. This groundbreaking battery, smaller than a coin, marks a significant advancement in the miniaturization of atomic energy. It's designed for various applications, including mobile phones and drones, potentially allowing these devices to operate indefinitely without charging. This development aligns with China's strategic goals under its 14th Five-Year Plan and represents a major leap in energy innovation.

For more details, you can read the full article on The Independent.

Advanced mathematics

📓 book

@programming_books_it

New Russian Radar Can Spot Tiniest Drones
https://sputnikglobe.com/20240622/new-russian-radar-can-spot-tiniest-drones-1119060515.html

Extract Musical Notes from Audio in Python with FFT - YouTube
https://www.youtube.com/watch?v=rj9NOiFLxWA

🤖 Musculature for Robots

A new type of “muscle” for robots has been developed, resembling human musculature. These rubber “muscles” can move efficiently thanks to individual servo motors, and the designers have also created special rotating shafts to facilitate their movement.

🤖@science #robots

🤖 Realistic Robo-Hand by Clone Robotics 🤖

Clone Robotics has developed a highly realistic robotic hand powered by electricity and heat. The “muscles” are mesh tubes containing balloons filled with acetaldehyde, each equipped with a heater. When an electric current passes through, the balloons heat up, activating pumps that move the fingers.

@science

Machine Learning: The Basics

📚 Book

@datascienceiot

76-page survey paper on Prompting Techniques ✨

Explores structured understanding and taxonomy of 58 text-only prompting techniques, and 40 techniques for other modalities.

📌 The paper focuses on discrete prefix prompts rather than cloze prompts, because prefix prompts are widely used with modern LLM architectures like decoder-only models. It excludes soft prompts and techniques using gradient-based updates.

📌 The paper identifies 58 text-based prompting techniques broken into 6 major categories:

1) In-Context Learning (ICL) - learning from exemplars/instructions in the prompt

2) Zero-Shot - prompting without exemplars

3) Thought Generation - prompting the LLM to articulate reasoning

4) Decomposition - breaking down complex problems

5) Ensembling - using multiple prompts and aggregating outputs

6) Self-Criticism - having the LLM critique its own outputs

📌 For ICL, it discusses key design decisions like exemplar quantity, ordering, label quality, format, and similarity that critically influence output quality. It also covers ICL techniques like K-Nearest Neighbor exemplar selection.

📌 Extends the taxonomy to multilingual prompts, discussing techniques like translate-first prompting and cross-lingual ICL. It also covers multimodal prompts spanning image, audio, video, segmentation, and 3D modalities.

📌 More complex techniques like agents that access external tools, code generation, and retrieval augmented generation are also taxonomized. Evaluation techniques using LLMs are discussed.

📌 Prompting issues like security (prompt hacking), overconfidence, biases, and ambiguity are highlighted. Two case studies - benchmarking techniques on MMLU and an entrapment detection prompt engineering exercise - are presented.

https://arxiv.org/abs/2406.06608

@opendatascience

Introduction to Applied Linear Algebra – Vectors, Matrices, and Least Squares

📚 Book

@datascienceiot

Forecasting in Economics, Business, Finance and Beyond

📚 Book

@datascienceiot

Fundamentals of Data Engineering

📌 Book
📌Download

@datascienceiot

Neural Networks and Deep Learning

📓 book

@datascienceiot

Collection of resources in the form of eBooks related to Data Science, Machine Learning, and similar topics

📖 Github

@datascienceiot