Tech News Review Week 1-9 March 2021, by Stelios Iordanis


Scientists use human cerebral organoid to test drug for deadly brain disease

The organoid model will be upgraded for screening of larger numbers of novel drug candidates and could one day beat Mad Cow disease (CJD) along with other neurodegenerative diseases.


Detecting hidden circulating tumor cells in non-small cell lung cancer patients

This is a novel method to better detect the circulating tumor cells (CTCs) that can show the existence of metastases in the case of Non-Small Cell Lung Cancer (NSCLC) .


Making the role of AI in medicine explainable

Pixel by pixel, these heatmaps highlight the AI decision process and the gravity of the region, thus providing plausibility of the results.


Why the lovable llama might be a secret weapon against COVID-19

Camelids such as llamas produce smaller antibodies, i.e. nanobodies ideal foe treatments against several diseases binding to areas of viral proteins that larger antibodies cannot fit into, blocking those proteins from connecting with cells.


Quantum computing

Xanadu announces programmable photonic quantum chip able to execute multiple algorithms

Xanadu claims to have found solutions to various problems associated with photonics-based systems and has come up with a working programmable photonic quantum chip that can run a multitude of algorithms in a scalable way.


Machine Learning


Assessing regulatory fairness through machine learning

Machine learning is being used to help managedecisions  that federal agencies are tasked to do—as a way to help increase efficiency, in a fair, transparent way.


dxaen/percival: Making In-Browser perceptual ad-blocking practical with Deep Learning

This is a browser-embedded, lighweight, deep learning powered ad blocker that can intercept rendering of iframes, images created by complex JavaScript transformations as well as Gifs and regular images. Percival analyzes these image frames and performs blocking based on deep learning image classification.


Beauty is in the brain: AI reads brain data, generates personally attractive images

Ultimately, the study could boost the capacity for computers to learn and increasingly understand subjective preferences, through interaction between AI algorithms and brain-computer interfaces and it could be applicable to other cognitive functions such as perception and decision-making in order to detect stereotypes or implicit bias and better model individual differences.


Google AI Blog: PAIRED: A New Multi-agent Approach for Adversarial Environment Generation

Creating training environments for machine learning and algorithms for this purpose is absolutely essential for the advancement of the field.


Self-supervised learning: The dark matter of intelligence

Facebook’s New AI Teaches Itself to See With Less Human Help | WIRED

Self-supervised learning, i.e. machine learning without using labeled examples is gaining traction and it will dominate the future progress of ML.

Facebook will release some of the technology behind Seer but not the algorithm itself because it was trained based on Instagram data corpus.



Student Wins James Dyson Award for Solar Panels Made from Food Waste

The new panels can generate energy almost half of the time, whereas current panels only produce 15-25% of the time. The production of bioluminescent particles used in the panels comes from a process of crushing the fruits and vegetables that would otherwise constitute a waste.


Breaking the warp barrier for faster-than-light travel

It remains to be seen if humanity will ever be able to minimize the astronomical amount of energy needed within the range of today’s technologies, such as a large modern nuclear fission power plant.