Ken Evans’ Post

Good news, my paper entitled "Implementation of the Gray Wolf Algorithm in Optimization of Artificial Neural Network Method for Fingerprint-Based Toxicity Prediction" has been officially released by IEEE at the 6th International Conference of Computer and Informatics Engineering (IC2IE). This research project focuses on detecting toxicity levels in chemical compounds that are the basic ingredients in making medicines using an Artificial Neural Network architecture that is optimized using the Grey Wolf Algorithm. I am very proud of the dedication and consistency with which I achieved this goal. This article is proof of my commitment to society to advance technology in the field of data science and provide a positive impact in the health sector. #IEEE #6thIC2IE #DeepLearning #DataScience #ArtificalNeuralNetwork #GreyWolfAlgorithm #Medicine #Research

The public artificial intelligence debate is often focused on the negative - great write up in The Economist discuss how Ai will revolutionize science - just like the Microscope “In 1665, during a period of rapid scientific progress, Robert Hooke, an English polymath, described the advent of new scientific instruments such as the microscope and telescope as “the adding of artificial organs to the natural”. They let researchers explore previously inaccessible realms and discover things in new ways, “with prodigious benefit to all sorts of useful knowledge”. For Hooke’s modern-day successors, the adding of artificial intelligence to the scientific toolkit is poised to do the same in the coming years—with similarly world-changing results.” https://lnkd.in/gRzGPyaz

Iliana Ivanova: "Artificial Intelligence (AI) is a powerful force for transformation. It also has the potential to be a major disruptor, with the potential to redefine the vast realm of science and to impact many aspects of human life. This seismic shift underscores the urgency for a tailored European research policy dedicated to AI in science." For more information - check UE Policy Brief: Harnessing the power of AI to accelerate discoverya nd foster innovation 👉 https://lnkd.in/dsyxMuf9

Dear friends, I am glad to share the recent recognition for contributions to the "Engineering Applications of Artificial Intelligence - EAAI," a journal by Elsevier/IFAC (International Federation of Automatic Control). In my role as a reviewer for multiple IEEE, Elsevier, and IFAC journals, I am dedicated to contributing to the advancement of science at the intersection of Engineering, Computer Science, and Artificial Intelligence. #ArtificialIntelligence (AI) plays a pivotal role in the ongoing Fourth and Fifth Industrial Revolution, and we are witnessing remarkable progress in various machine learning methodologies. #AI techniques are widely adopted by engineers to tackle a diverse array of challenges. Let's continue to shape the future of technology and innovation. [Link to the journal: https://lnkd.in/dgZgKhPb] #ArtificialIntelligence #MachineLearning #ScientificContributions #ResearchRecognition #EngineeringExcellence #TechInnovation #ScienceCommunity

There's still time to submit an abstract to AAIC! As an added incentive - if you are a student or postdoc, please answer AI PIA to the question "Please select the Professional Interest Area this submission is most closely aligned with" to have your abstract considered for a AI PIA Poster Award!

Numbers and measurements speak for themselves

Day 16 of the #100DaysAmplifierDesign.   Today's learning took me towards the use of common-source amplifiers in artificial neural networks.  In a research paper I found online titled "Common-Source Amplifier Based Analog Artificial Neural Network Classifier" by Akshay Jayaraj and Imon Banerjee. The authors presented an innovative approach to analog artificial neural networks (ANNs) using common amplifiers. Here are the key points from the paper: Objective: The authors aim to design an ANN classifier using a common-source amplifier-based nonlinear activation function.   Methodology: - They created a shallow ANN using transistor-level circuits.  - The proposed ANN classifier leverages a common-source amplifier as a key component. - The activation function is designed to be nonlinear, allowing for complex mappings.   Performance: - The classifier is evaluated based on the MNIST dataset, which consists of handwritten digits. - Achieved a multinomial (10 classes) classification accuracy of 0.82.   Significance: - Analog ANNs have potential applications in low-power, hardware-efficient machine learning systems. - This work demonstrates the feasibility of using analog components for neural network computation.   In summary, the paper introduced a novel approach to analog neural network design, emphasizing the use of common-source amplifiers and achieving promising classification accuracy on the MNIST dataset.   I have linked the paper below if you are interested in it. https://lnkd.in/dXpaGc_R Stay tuned for more updates as I continue my #100DaysAmplifierDesign journey.  Pipeloluwa Olayiwola TSMC Tiny Tapeout NVIDIA IEEE Solid-State Circuits Society

🤖 Project: Building like Gemini ai Telegram Bot In this project, I developed a Telegram bot  AIMER Society - Artificial Intelligence Medical and Engineering Researchers Society

Explore research paper on 'Enhancing Transformer Models with Abacus Embeddings for Superior Arithmetic and Algorithmic Reasoning Performance' In a recent study, researchers from the University of Maryland, Lawrence Livermore National Laboratory, Tübingen AI Center, and Carnegie Mellon University introduced a novel method called Abacus Embeddings. This approach significantly enhances the transformer model’s ability to track the position of each digit within a number. Abacus Embeddings assign the same positional embedding to all digits of the same significance, enabling the model to align digits correctly. The Abacus Embeddings technique combines positional embeddings with input injection and looped transformer architectures. By encoding the relative position of each digit within a number, the model can more accurately perform arithmetic operations. For instance, the researchers trained transformer models on addition problems involving up to 20-digit numbers and achieved up to 99% accuracy on 100-digit addition problems. This represents a state-of-the-art performance, significantly surpassing previous methods. Quick read:https://lnkd.in/gDHzDhCk Paper: arxiv.org/abs/2405.17399 GitHub: https://lnkd.in/gddWeR34

🕵️ Think your fingerprints are one-of-a-kind? Think again! Columbia Engineering's undergrads, led by Gabe Guo, are changing the game in forensic science. They've used AI to show that fingerprints from the same person can be similar, challenging old beliefs. Their AI analyzed 60,000 fingerprints and found a 77% accuracy rate, a huge step up for forensic science. Initially met with skepticism, their findings are now published in Science Advances. This isn't just a win for forensics – it's a showcase of AI's power to revolutionize science! Source: Columbia Engineering