CTrees’ Post

Our paper introduces a pioneering hybrid methodology for long-term vegetation prediction, combining the Enhanced Vegetation Index (EVI), the Nonlinear AutoRegressive with eXogenous inputs Artificial Neural Network (NARX ANN) algorithm, and wavelet analysis. This approach demonstrates remarkable robustness in forecasting future vegetation changes, even in the face of environmental variability. The paper is freely accessible, and you can find it [https://lnkd.in/dnSrrRxQ].

As I keep learning why different scientific/Algorithmic techniques work, I also actively learn their history. Here's what I stumbled upon when I was studying the history of #Btrees. Previously when I was going through the history of #ANN Artificial Neural Networks I learnt that it came out of Cornell's Aerospace lab. Now the most prominent data structure running most databases came out of a similar space. #computerscience #database #storage #datastructures #boeing #btrees #indexes #history

The field of imageomics aims to help explore fundamental questions about biological processes on Earth by combining images of living organisms with computer-enabled analysis and discovery. #Optics #MachineVision

As we have spoken about in great detail on MLST over the years, basic NN architectures have severe computational limitations. They are not Turing machines, they can not perform the types of computation a simple calculator can perform. But what if we could have our cake and eat it? The future might just be neurosymbolic! Dr. Petar Veličković is a rising star scientist at DeepMind and is taking his Geometric Deep Learning framework to the next level in collaboration with his collegues. This is just a teaser clip but there will be lots more info coming.

Regression problems are more challenging than classification problems for deep learning (DL) to solve. Here, we're indeed applying DL for a complex regression task. This is a long-lasting (so far three years) project to tackle an urgent problem: peak ground acceleration (PGA) prediction using a very short window of seismograms (just like predicting the stock price but with fewer historical observations), which is critical to the reliability of the on-site alert and, thus, the success of earthquake early warning (EEW). Like many others, we initially doubted whether DL could play a role in such a conundrum in earthquake seismology. Luckily enough, we made it work after many failed trials. The work is finally published! The high prediction accuracy of PGA using very little time (as small as 2 seconds) after P-wave arrival enables an immediate but high-fidelity on-site alert right before (several more seconds) the most destructive wave hits the ground after a devastating earthquake occurs. Note that, in EEW, every second matters! Many thanks to this incredible team!

🚨 New article in PNAS News 🚨 with Begüm Çeliktutan and Romain Cadario! https://lnkd.in/egrB4WrQ We find that people see more of their biases in algorithms' decisions than in their own decisions, even when algorithms are trained on their decisions and when those decisions are the same. Algorithms seem to reveal our bias blind spot. People saw more bias in ratings attributed to algorithms, even algorithms trained on their own ratings, as in ratings attributed to others. Perhaps most exciting is that because people see more #bias in #algorithms than in themselves, we find they are also more willing to correct for bias in algorithms' decisions than in their own decisions. We hope our work stimulates work that uses algorithms as a tool for debiasing.

Together with Konstantina Safouri and Georgios Varsamis we had the opportunity to attend the prestigious Big Data from Space 2023 conference in Vienna, where we presented our recent work, titled " Hybrid Classical - Quantum Neural Network for Rice Crop Identification". During the conference we had the pleasure to attend a number of interesting talks and interact with professionals from different backgrounds and gained useful insights. We would like to thank the organising committee of the #BiDS_2023 conference for this wonderful event. #Bigdatafromspace #quantumcomputing #ml #dataanalysis #spacedata #BiDS2023

Hi! Here is my first talk at the University of Potsdam in the Seminar of AI in Software Engineering. The presentation is based in the article "Is Deep Learning Good Enough in SDP?". The focus is on two CNN models: SqueezeNet and Bottleneck. The authors conducted a detailed comparative study across seven datasets from the NASA repository, comparing the performance of these models against baseline models. Paper's link: https://lnkd.in/e88-rp_v

The world isn't going that well, is it? There's not much intelligent to say, at least for me. I've then closed up in my small dungeon and wrote a new piece about entropy and how you can use it in Machine Learning. https://lnkd.in/eErKZ--p #datascience #machinelearning #entropy

I'm so excited to share the new paper of the I-FENN series (Integrated Finite Element Neural Network) with non-local gradient damage propagation. So far, the feasibility of our framework was shown on a single load-increment basis, and two crucial questions were lingering: Can we model the entire load-history of damage propagation with I-FENN, and is it indeed faster than conventional FEM? Here, for the first time, we address both questions. Check the answers in the link below! Mostafa Mobasher Habiba Eldababy @Diab Abueidda