Kenza Bouzid

👉🏼 Detection of ChatGPT fake science with the xFakeSci learning algorithm 🤓 Ahmed Abdeen Hamed 👇🏻 https://lnkd.in/eWXP83PT 🔍 Focus on data insights: - The study demonstrates how AI-generated content can be distinguished from scientific articles. - Introduction of xFakeSci, a learning algorithm capable of identifying ChatGPT-generated articles. - Training and calibration steps incorporated to enhance algorithm performance. 💡 Main outcomes and implications: - xFakeSci algorithm achieved F1 scores ranging from 80 to 94%, outperforming traditional data mining algorithms. - Calibration and proximity distance heuristic contribute to the algorithm's promising performance. - Significant step towards combating fake science generated by AI tools like ChatGPT. 📚 Field significance: - AI-generated content detection - Data-driven heuristics - Combating fake science 🗄️: [#AI #datainsights #fakeScience #xFakeSci #ChatGPT]

1

Kolmogorov–Arnold Networks seem like a paradigm shift in ML architectures. Curious to see how they evolve. "While MLPs have fixed activation functions on nodes ("neurons"), KANs have learnable activation functions on edges ("weights"). KANs have no linear weights at all -- every weight parameter is replaced by a univariate function parametrized as a spline. " https://lnkd.in/gdPk_qCQ

29

Learn how you can convert a PyTorch model, trained on the web, to a mobile-optimized format. Madhur Zanwar describes the process using a few lines of code. He also discusses his journey and how one approach didn't yield satisfactory results while trying to run it in a React Native app. #machinelearning #pytorch #computervision #edgeml #reactnative #objectdetection #yolov5

14

Large Generative Models (LGMs) such as GPT and Stable Diffusion have significantly enhanced the efficiency of real-world applications by producing creative and meaningful content. The success of these models is attributed to extensive training on well-curated datasets spanning a wide range of fields, in contrast to previous models that were limited to small datasets for specific tasks. Researchers have now introduced the Large Graph Generative Model (LGGM), which surpasses other graph generative models in zero-shot capabilities and can be fine-tuned with domain-specific graphs. LGGM can generate graphs based on text prompts, offering users detailed control over the created graphs. Training LGGM requires a vast corpus of graphs from various domains, sourced from the Network Repository across 13 different fields. When tested on the same graphs used for training, LGGM outperformed DiGress (a diffusion model), particularly in situations with fewer available graphs. This advantage is beneficial for semi-supervised settings such as anomaly detection software and drug design. Future research will focus on evaluating the practical applications and usefulness of LGGM in various real-world scenarios. https://lnkd.in/eHHxe5Nj

4

🚀 Want to employ the power of #PyTorch for your scientific models without a steep learning curve? 👀 Check out our latest blog, where Travis Oliphant discusses how PyTorch's integration with the Array API makes it easy to use with familiar libraries like NumPy and scikit-learn.🔥 https://lnkd.in/dDRkEY3v #Innovation

10

𝗘𝘅𝗰𝗶𝘁𝗶𝗻𝗴 𝗡𝗲𝘄𝘀 𝘄𝗶𝘁𝗵𝗶𝗻 𝘁𝗵𝗲 𝗟𝗮𝗿𝗴𝗲 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗠𝗼𝗱𝗲𝗹𝘀 (𝗟𝗟𝗠𝘀) 𝗦𝗽𝗮𝗰𝗲! PyTorch has officially released TorchTune, a new tool designed to fine-tune Large Language Models (LLMs) with ease and efficiency. This groundbreaking tool promises to streamline the fine-tuning process, making it more accessible to researchers and practitioners alike. https://lnkd.in/eMf6dNZJ 𝗧𝗼𝗿𝗰𝗵𝘁𝘂𝗻𝗲 𝗢𝗳𝗳𝗲𝗿𝘀: ▪ 𝗣𝘆𝗧𝗼𝗿𝗰𝗵-𝗻𝗮𝘁𝗶𝘃𝗲 𝗟𝗟𝗠 𝗶𝗺𝗽𝗹𝗲𝗺𝗲𝗻𝘁𝗮𝘁𝗶𝗼𝗻𝘀: Build on familiar ground with composable building blocks for popular LLMs. ️ ▪ 𝗠𝗲𝗺𝗼𝗿𝘆-𝗲𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝘁 𝗳𝗶𝗻𝗲-𝘁𝘂𝗻𝗶𝗻𝗴 𝗿𝗲𝗰𝗶𝗽𝗲𝘀: Train on consumer-grade GPUs with 24GB VRAM – no beefy machines needed! ▪ 𝗘𝘅𝘁𝗲𝗻𝘀𝗶𝗯𝗶𝗹𝗶𝘁𝘆 𝗮𝗻𝗱 𝗲𝗮𝘀𝗲 𝗼𝗳 𝘂𝘀𝗲:  YAML configs and support for common dataset formats streamline your workflow. 𝗙𝗼𝗿 𝘁𝗵𝗼𝘀𝗲 𝗶𝗻𝘁𝗲𝗿𝗲𝘀𝘁𝗲𝗱 𝗶𝗻 𝗱𝗶𝘃𝗶𝗻𝗴 𝗱𝗲𝗲𝗽𝗲𝗿, 𝗵𝗲𝗿𝗲 𝗮𝗿𝗲 𝘀𝗼𝗺𝗲 𝗿𝗲𝘀𝗼𝘂𝗿𝗰𝗲𝘀: Code: https://lnkd.in/gEgETe5t Blog: https://lnkd.in/gP3mctA6 Tutorials: https://lnkd.in/gvh2eEsU This is a fantastic opportunity to explore and innovate, I'd be very interested in hearing your thoughts. ▪ 𝗪𝗵𝗮𝘁’𝘀 𝘆𝗼𝘂𝗿 𝘁𝗮𝗸𝗲 𝗼𝗻 𝘁𝗵𝗶𝘀?  ▪ 𝗛𝗼𝘄 𝗱𝗼 𝘆𝗼𝘂 𝘀𝗲𝗲 𝗧𝗼𝗿𝗰𝗵𝗧𝘂𝗻𝗲 𝗶𝗺𝗽𝗮𝗰𝘁𝗶𝗻𝗴 𝘆𝗼𝘂𝗿 𝗰𝘂𝗿𝗿𝗲𝗻𝘁 𝗽𝗿𝗼𝗷𝗲𝗰𝘁𝘀 𝗼𝗿 𝗿𝗲𝘀𝗲𝗮𝗿𝗰𝗵? ▪ 𝗛𝗼𝘄 𝗱𝗼 𝘆𝗼𝘂 𝘁𝗵𝗶𝗻𝗸 𝗧𝗼𝗿𝗰𝗵𝗧𝘂𝗻𝗲 𝘄𝗶𝗹𝗹 𝗶𝗺𝗽𝗮𝗰𝘁 𝘁𝗵𝗲 𝗳𝘂𝘁𝘂𝗿𝗲 𝗼𝗳 𝗺𝗮𝗰𝗵𝗶𝗻𝗲 𝗹𝗲𝗮𝗿𝗻𝗶𝗻𝗴 𝗮𝗻𝗱 𝗔𝗜 𝗱𝗲𝘃𝗲𝗹𝗼𝗽𝗺𝗲𝗻𝘁? #AI #PyTorch #TorchTune #LLMs #MachineLearning

24

1 Comment

I'm excited to share my latest article on "What’s Behind the Architecture of FAISS in Semantic Search"! In this deep dive, I explore the inner workings of Facebook AI Similarity Search (FAISS) and its role in powering semantic search, a key innovation driving modern AI and machine learning applications. From vector embeddings to efficient similarity retrieval, this article breaks down the core concepts, architecture, and use cases of FAISS. Check it out to understand how FAISS makes large-scale search faster and smarter! 🔗 #AI #SemanticSearch #FAISS #MachineLearning #ArtificialIntelligence #NLP #DataScience #TechInnovation Pooja Porchezhian

28

Combining the power of TensorFlow, NeuralForecast, and zero-shot LLMs, Jinhang Jiang shows how we can generate accurate multi-step quantile forecasts.

11

OpenAI has just released the Multilingual Massive Multitask Language Understanding (MMMLU) dataset on Hugging Face, providing a valuable resource for evaluating multilingual large language models (LLMs). The MMLU test set is available in 14 languages, including Arabic, German, Spanish, and French. It spans 57 categories, ranging from elementary to advanced professional subjects, and has been translated by professional human translators. This dataset allows for the evaluation of AI models' general knowledge across diverse cultures and is used in the openai/simple-evals project. However, the licensing details remain unclear at this time. Explore the dataset: https://lnkd.in/er9iPrNA Kudos to OpenAI for this significant contribution! #OpenAI #MultilingualAI #LLM #NaturalLanguageProcessing #MMLU #AIResearch #HuggingFace #DatasetRelease #LanguageModels #AITranslation #DeepLearning #AIInnovation #GlobalAI #MachineLearning #AIEvaluation Umar Iftikhar

3

Excited to share my latest article on Medium: "Computation of LLMs: CPU vs GPU - A Detailed Analysis". Dive into the nuances of how the choice of hardware impacts the performance of large language models (LLMs). Discover why CPUs excel with low-dimensional data while GPUs take the lead as data dimensionality increases. Whether you're a machine learning enthusiast or a seasoned professional, this comprehensive guide will help you optimize your computational resources effectively. 🔗 Read the full article here and let me know your thoughts! #MachineLearning #AI #DataScience #CPUvsGPU #LLMs #DeepLearning #NLP #TechInsights #gpu #cpu #computation #llama3 #llm #benchmarking #ai #machinelearning #deeplearning #openai #genai Pooja Porchezhian

37

OpenAI has just unveiled its latest state-of-the-art reasoning model, and it’s a game-changer! 🌟🤖 This new model not only sets a new standard in artificial intelligence but excels in areas previously thought challenging for AI. 🔬 Key Highlights: 1. PhD-Level Physics Mastery: The model has demonstrated an impressive ability to handle complex PhD-level physics problems with unparalleled accuracy and depth. This breakthrough could revolutionize how we approach advanced scientific research and education. 2. Olympiad Performance: It’s not just about theory—this model excels in practical applications too! It has shown exceptional performance in international Olympiads, outshining previous models and showcasing its prowess in solving high-level competitive problems. 3. Efficient Code Writing: The new model writes code with greater efficiency and clarity, making it a valuable asset for developers and engineers seeking to streamline their coding practices. 4. Top-Tier Contest Performance: Impressively, it ranks in the 89th percentile in Codeforces contests, highlighting its capability to tackle competitive programming challenges with agility and skill. This advancement underscores the remarkable progress in AI, pushing the boundaries of what’s possible in​ the world of Tech, Science, and beyond! Written using ChatGPT :) (jk) https://lnkd.in/gJXMHHdT

5

Is GPT-4o good at table recognition? Not quite! OCR looks fine, but tripped by spanning cell table structure! 🤩 However, our UniTable, with only 124M param trained by self-supervision, shows great understanding! 👉 UniTable paper: arxiv.org/abs/2403.04822 👉 UniTable code: https://lnkd.in/e4a7YMxJ #MachineLearning #GPT-4o #Table #Document

85

8 Comments

Check out this course from deeplearning.ai, explaining the advanced capability of function calling in LLMs. #LLMs #GAI #NexusAI

4