TEDxPortland’s Post

https://lnkd.in/gutp-dg7 "Project name: Quantum computing(Grover's algorithm) by Artificial intelligence" Contributor: Rajeev Singh Sisodiya Quantum computing, a revolutionary paradigm in computation, harnesses the principles of quantum mechanics to perform computations that were once thought to be impossible for classical computers. Grover's algorithm, a prominent quantum algorithm, stands out for its potential to significantly speed up unstructured search problems. This paper explores the application of artificial intelligence (AI) techniques in the understanding and optimization of Grover's algorithm. Quantum computing and artificial intelligence (AI) are two cutting-edge technologies that have the potential to revolutionize various fields. While they are distinct fields, the intersection of quantum computing and AI holds promise for solving complex problems more efficiently than classical methods. while quantum computing is still in its early stages of development, it has the potential to impact various aspects of artificial intelligence. Researchers and practitioners are actively working on harnessing the unique capabilities of quantum computers to enhance AI algorithms and solve problems that are currently intractable for classical computers.

Sundar Pichai revealed that over 25% of Google’s code is now written by AI. Is this the start of a new evolution or the eventual extinction of software engineers?

https://lnkd.in/gPMRQHew "Project Name: Google-Quantum AI with Introduction to CIRQ" Contributor: Rajeev Singh Sisodiya CIRQ: Cirq is a framework for writing quantum algorithms for noisy intermediate scale quantum (NISQ) devices. Roughly speaking, NISQ devices are those with O(100) qubits that can enact O(1000) gates. Because the resources for NISQ devices are so constrained, we believe that a framework for writing programs on these devices needs to be aware of all of the architectural properties of the device on which the algorithm is written. This is in contrast to other frameworks where there is a clean separation between the abstract model being used and the details of the device.

🌈 Research Update: Aligning AI through Quantum Operations By integrating classical and quantum computing, I managed to achieve excellent alignment results, demonstrating the powerful synergy between these paradigms. With this method I apply Inverse Reinforcement Learning, for real-time error-proof focus, aiming to enhance any AI model. I refined and made my architecture more elegant, but also meaningful - now directly reflecting the complete anatomy of biological neurons (12 parts of them). Stay tuned as I continue to refine this model and explore its vast potential across various industries. Your thoughts and feedback on this intersection of AI and quantum computing are highly welcome! #QuantumComputing #ArtificialIntelligence #Innovation #Technology #DataScience #MachineLearning #AIResearch #Qiskit #IBMQuantum You may play with the full code here: https://lnkd.in/d8b87xeR Documentation: https://lnkd.in/daEuCSiy

Thrilled to share that I recently participated and won 3rd in the Prompt Olympics, an exciting event where Amazon Web Services (AWS) teamed up with AI Thinkerers at the MoneyLion TRX office. The event featured two thrilling challenges aimed at mastering prompt engineering: 1. LLM Lockpicking - Crafting the perfect questions to get the LLM Password Protector to reveal secret codes and advance to the next levels. 2. Word Relay - Using creative descriptions to help your LLM Relay Partner guess difficult words – with a twist! Coming from a product design background, this experience was not only fun but also eye-opening. I’m excited to explore how prompt engineering can further enhance my career. I can't wait to see where this journey takes me! #PromptEngineering #AI #AWS #ProductDesign #LearningAndGrowth

With their research heritage (from both Brain and Deep Mind), their vast revenue resources, and their self-investment in efficient compute, Google was always going to re-establish their position as an AI front runner. Amazing to see how they have ended the year (Gemini 2, Veo2, Imagen 3). But can they go beyond their undisputed engineering prowess and actually reshape their products and business around it? Or has their search dominance corroded their ability to create new things that people want, leading a lot of people to accuse them of being stuck, old, unable to change?

👋 Hello again, Information Extraction enthusiasts! 👋 After introducing ReLiK the other day, we received a question: "Does it work on Colab?" 🤔 This got us thinking—if we’re truly committed to delivering an "academic-budget" IE model, Colab Free GPU should be the baseline hardware we aim to support. Initially, our previous index (with over 6M entities) used an e5 base retriever, making the embeddings too large to fit on Colab. But we’ve got great news—ReLiK now works seamlessly on Colab! 🎉 We’ve developed a new configuration for a tiny cIE model that performs both Entity Linking and Relation Extraction using just 5.6GB of GPU RAM and 5.3GB of system RAM (with a peak of 9GB, as the index loads on the CPU before moving to the GPU). This is all made possible by a new retriever using e5 small, which can be loaded on bf16 and takes up only about 4GB of GPU RAM. We’re excited to announce these two new retrievers—base and small—that are already integrated into our latest configs: - https://lnkd.in/dJUs5fbw (small retriever + small reader) - https://lnkd.in/dsx-nQgu (base retriever + small reader) These models are trained with matryoshka loss, allowing them to be trimmed down while still outperforming our previous base retriever. There are some other new tricks to alleviate the memory usage which I will let Riccardo Orlando explain. Try it out for yourself on Colab: https://lnkd.in/dC2V6Yjz #NLP #EntityLinking #RelationExtraction #AI #MachineLearning #InformationExtraction #ReLiK #Colab #SapienzaNLP #Research #Innovation

💡 The use of Large Language Models (#LLMs) as judges to assess and compare different LLMs has garnered significant attention in recent research. This innovative method provides a scalable and objective approach to benchmarking models, improving our ability to evaluate their strengths and weaknesses effectively. 🌟 In my latest notebook, I explored this concept by comparing two prominent models: - #Model A: Google/gemma-2-2b-it - #Model B: Meta-llama/Llama-3.2-1b-instruct 🌟 Additionally, a third model, NVIDIA/Mistral-NeMo-Minitron-8B-Instruct, acts as the judge to evaluate and determine the model that offers better responses based on predefined criteria. 🔗 Explore the notebook here:👇 (https://lnkd.in/gaCQ5eWp).

AI represents a fundamental shift in how we think of software programming. Those who win in this era will share a common characteristic, they have figured out how to efficiently harness the power of both Data and Scale. I wrote this piece "The Joy (and Challenges) of Brute Force Computing" to give a perspective on the fundamentals of this shift in approach to computing that is now well underway in the leading companies. https://lnkd.in/eFdjfVQm Its a quick 4 minute read. Please share your thoughts in the comments, i read them all.

Red Hat, the world's leading provider of open source solutions, has announced the completion of its acquisition of Neural Magic, a pioneer in software and algorithms designed…