Karim Guirguis

✨ Exciting insights from Yann Dubois's recent lecture at Stanford University on building Large Language Models (#LLMs) 🟢 Components of LLMs: Focus on architecture, training algorithms, data collection, evaluation methods, and system optimization for efficient model deployment. 🟢 Pre-training vs. Post-training: Understand the classical language modeling paradigm and the emerging trend of fine-tuning models like ChatGPT to function as AI assistants. 🟢 Importance of Tokenization: Effective tokenization strategies are crucial for handling diverse languages and ensuring model performance. 💎 This lecture is a must-watch for anyone interested in #AI advancements! 👉 https://lnkd.in/dcevp3cJ

44

1 Kommentar

✨ 𝐃𝐢𝐯𝐢𝐧𝐠 𝐃𝐞𝐞𝐩 𝐢𝐧𝐭𝐨 𝐇𝐮𝐠𝐠𝐢𝐧𝐠 𝐅𝐚𝐜𝐞 𝐰𝐢𝐭𝐡 𝐒𝐭𝐚𝐧𝐟𝐨𝐫𝐝'𝐬 𝐂𝐒𝟐𝟐𝟒𝐍 𝐂𝐨𝐮𝐫𝐬𝐞 🟢 Hugging Face Overview: Explored the powerful tools and libraries offered by Hugging Face for NLP projects. 🟢 Model Implementation: Hands-on experience implementing state-of-the-art models like BERT and GPT using Hugging Face. 🟢 Practical Applications: Applied Hugging Face models to real-world NLP tasks such as sentiment analysis and machine translation. 🟢 Community and Resources: Discovered the vibrant Hugging Face community and the resources available for continuous learning. 💡 Feeling empowered and ready to leverage Hugging Face for innovative NLP solutions! 👉 https://lnkd.in/dq9AePqR #NLP #DeepLearning #HuggingFace #StanfordCS224N #AI #MachineLearning

12

1 Kommentar

✨ Exciting developments at the intersection of Large Language Models (#LLMs) and Reinforcement Learning (#RL)! 🚀 Let's see how pre-trained LLMs can be leveraged to enhance RL in several key ways: 🔹 LLMs for Pre-processing Input: If the input observations sent to the agent are in natural language, it is natural to use an LLM to process them, in order to compute a more compact representation. 🔹 LLMs for Reward Functions: Designing reward functions can be challenging, but LLMs can help. For example, LLMs can provide preference judgements (like a human would) for RLHF (Reinforcement Learning from Human Feedback) systems, thereby generating training data for a reward model. 🔹 LLMs for World Models: Instead of learning world models from scratch, pre-trained LLMs can serve as world models. For example, code as a reward. Where prompt a VLM with an initial and goal image, and ask it to describe the corresponding sequence of tasks needed to reach the goal. Then ask the LLM to synthesize code that checks for completion of each subtask. These reward functions are then “verified” by applying them to an offline set of expert and random trajectories; a good reward function should allocate a high reward to the expert trajectories and a low reward to the random ones. Finally, the reward functions are used as auxiliary rewards inside an RL agent. 🔹 LLMs for world models: There are many papers that use transformers or diffusion models to represent the world model, and learn them from data collected by the RL agent. 🔹 LLMs for Policies: LLMs can be used as policies, either by pre-training on state-action sequences or by using in-context learning with prompts. In-context learning allows LLMs to "reason" step-by-step before predicting an action, and this can be enhanced by retrieving relevant past examples. ⚡ In summary, LLMs are proving to be versatile tools in RL, offering benefits from input processing to policy generation. This is a rapidly evolving field with much potential, stay tuned for more! 👉 https://lnkd.in/dRKRem9P #ReinforcementLearning #LargeLanguageModels #AI #MachineLearning

24

I will be speaking at the AIAI Berlin conference on the topic of Automating Computer Vision Workflows with Foundation Models by AI Accelerator Institute along with Mirza Rahim Baig at the Waldorf Astoria on 11th and 12th. AI Accelerator Institute This is one of the biggest AI summits in Germany and I am excited to be there See you there! #artificialintelligence #businessintelligence #leadership #genai #computervision

60

3 Kommentare

💎 𝐊𝐧𝐨𝐰𝐢𝐧𝐠 𝐖𝐡𝐞𝐧 𝐭𝐨 𝐀𝐬𝐤 from Google Large Language Models (#LLMs) are prone to generating factually incorrect information when responding to queries that involve numerical and statistical data or other timely facts. 💡 This paper presents an approach for enhancing the accuracy of LLMs by integrating them with Data Commons, a vast, open-source repository of public statistics from trusted organizations like the United Nations (UN), Center for Disease Control and Prevention (CDC) and global census bureaus. ⚡ To primary methods are explored: 1️⃣ Retrieval Interleaved Generation (#RIG), where the LLM is trained to produce natural language queries to retrieve data from Data Commons, 2️⃣ Retrieval Augmented Generation (#RAG), where relevant data tables are fetched from Data Commons and used to augment the LLM's prompt. 🚀 These methods are evaluated on a diverse set of queries, demonstrating their effectiveness in improving the factual accuracy of LLM outputs. 🚀 This work represents an early step towards building more trustworthy and reliable LLMs that are grounded in verifiable statistical data and capable of complex factual reasoning. 👉 Paper https://lnkd.in/daSwTMBz ♻️𝒌𝒊𝒏𝒅𝒍𝒚 𝒓𝒆𝒑𝒐𝒔𝒕♻️

17

Amazing conversation with Dr. Serrano in today's episode. I'd like to learn more from his book!

1

🧠 𝗘𝘃𝗲𝗿 𝘄𝗼𝗻𝗱𝗲𝗿𝗲𝗱 𝗵𝗼𝘄 𝗔𝗜 𝗰𝗼𝘂𝗹𝗱 𝗿𝗲𝘃𝗼𝗹𝘂𝘁𝗶𝗼𝗻𝗶𝘇𝗲 𝘀𝗰𝗶𝗲𝗻𝘁𝗶𝗳𝗶𝗰 𝗱𝗶𝘀𝗰𝗼𝘃𝗲𝗿𝘆? This new paper on 𝗔𝗜-𝗚𝗲𝗻𝗲𝗿𝗮𝘁𝗲𝗱 𝗦𝗰𝗶𝗲𝗻𝗰𝗲 (𝗔𝗜𝗚𝗦) explores how AI can autonomously complete the entire research process, from ideation to discovery. 💡 The paper argues that 𝗳𝗮𝗹𝘀𝗶𝗳𝗶𝗰𝗮𝘁𝗶𝗼𝗻—testing and disproving hypotheses—is essential for both human and AI-driven research. Current systems either lack this component or rely too heavily on specialized verification engines, limiting their applications. To address this, the study introduces 𝗕𝗮𝗯𝘆-𝗔𝗜𝗚𝗦, a multi-agent system where different agents handle key research stages, including a FalsificationAgent that verifies potential discoveries. 💡 Experiments show that Baby-AIGS can produce meaningful scientific findings, though not yet at the level of experienced human researchers. 💬 I see tremendous potential in AIGS to accelerate scientific breakthroughs. However, we must also consider the ethical implications and limitations of such systems. 👉 Paper https://lnkd.in/d46druRN #AI #Innovation #ScientificResearch

1

" Madhu Sudan (1966-present): The Algorithmic Architect 🖥 Who revolutionized the world of computer science with his groundbreaking work on algorithms, forever changing our understanding of computational complexity. As a pioneering theorist, Sudan made significant contributions to the development of probabilistically checkable proofs (PCPs), error-correcting codes, and pseudorandomness. His work laid the foundation for advancements in cryptography, coding theory, and machine learning. But Sudan's curiosity didn't stop at algorithms. He also made notable contributions to the study of pseudorandomness, property testing, and derandomization. By bridging gaps between theory and practice, he opened doors to new possibilities in computer science. Today, his work continues to inspire generations of researchers, proving that one person can indeed shape the future of technology!" #probability #research #algorithm #architect #complexity #PCPs #error #codes #cryptography #pseudorandomeness

5

🚨SIGGRAPH 2024 Paper Alert 🚨 ➡️Paper Title: CWF: Consolidating Weak Features in High Quality Mesh Simplification 🌟Few pointers from the paper 🎯In mesh simplification, common requirements like accuracy, triangle quality, and feature alignment are often considered as a trade-off. Existing algorithms concentrate on just one or a few specific aspects of these requirements. 🎯For example, the well-known Quadric Error Metrics (QEM) approach prioritizes accuracy and can preserve strong feature lines/points as well, but falls short in ensuring high triangle quality and may degrade weak features that are not as distinctive as strong ones. 🎯In this paper, Authors have proposed a smooth functional that simultaneously considers all of these requirements. The functional comprises a normal anisotropy term and a Centroidal Voronoi Tessellation (CVT) energy term, with the variables being a set of movable points lying on the surface. 🎯The former inherits the spirit of QEM but operates in a continuous setting, while the latter encourages even point distribution, allowing various surface metrics. 🎯Authors further introduced a decaying weight to automatically balance the two terms. They selected 100 CAD models from the ABC dataset, along with 21 organic models, to compare the existing mesh simplification algorithms with theirs. 🎯 Experimental results reveal an important observation: the introduction of a decaying weight effectively reduces the conflict between the two terms and enables the alignment of weak features. 🎯This distinctive feature sets their approach apart from most existing mesh simplification methods and demonstrates significant potential in shape understanding. 🏢Organization: Shandong University, The University of Texas at Dallas,  Wayne State University, Qingdao University of Science and Technology,  The University of Hong Kong, Texas A&M University 🧙Paper Authors: Rui Xu, Longdu Liu, Ningna Wang, Shuangmin Chen, Shiqing Xin, Xiaohu Guo, Zichun Zhong, Taku Komura, Wenping Wang , Changhe Tu 1️⃣Read the Full Paper here: https://lnkd.in/gu-KVMpm 2️⃣Project Page: https://lnkd.in/guii4W2W 3️⃣Code: https://lnkd.in/gB5Jq-aM 4️⃣Blender Plugin: https://lnkd.in/grgYH7_h 🎥 Be sure to watch the attached Demo Video -Sound on 🔊🔊 Find this Valuable 💎 ? ♻️REPOST and teach your network something new Follow me 👣, Naveen Manwani, for the latest updates on Tech and AI-related news, insightful research papers, and exciting announcements. #SIGGRAPH2024

15

The Godfather of AI is at #VivaTech ! Yann LeCun (@ylecun) advises students coming into the industry: "Don't work on LLM. This is in the hands of large companies, there's nothing you can bring to the table. You should work on next-gen AI systems that lift the limitations of LLMs.

1

Excited to share our latest paper titled "Hierarchical SVM for Semantic Segmentation of 3D Point Clouds for Infrastructure Scenes" by Mohamed Mansour, Jan Martens and Jörg Blankenbach! 📄 The paper addresses the integration of Building Information Modeling (BIM) and advanced 3D point cloud technologies. The focus lies on utilizing supervised machine learning to optimize the segmentation of 3D point clouds in infrastructure scenes with the aim of minimizing manual work in the creation of digital 3D models. The findings provide valuable insights for enhancing the efficiency and sustainability of civil engineering projects. Full Paper: https://lnkd.in/ez7yhKg7   #Research #BIM #MachineLearning #3DModeling #Infrastructure

18

Very interesting article about using #ChatGPT in #Robotics. #LLM

14

“Nougat can easily parse PDF-format academic documents, extract mathematical formulas and tables, and convert them into structured data for subsequent processing and analysis.“ Read “Advanced RAG Retrieval Strategy: Embedded Tables“ by zhaozhiming on Medium: https://lnkd.in/dFBUMW3n

4

> Characterizing Out-of-Distribution Error via Optimal Transport Yuzhe Lu, Yilong Qin, Runtian Zhai, Andrew Shen, Ketong Chen, Zhenlin Wang, Soheil Kolouri, Simon Stepputtis, Joseph Campbell, Katia Sycara This strikes me as a novel way of bounding error across contexts. https://lnkd.in/eMXB-NkK

4

The next big thing that may replace diffusion in the field of image generation may be coming -- next-scale prediction. This paper, "Visual Autoregressive Modeling: Scalable Image Generation via Next-Scale Prediction", is probably the most exciting paper I've read so far this year. Like always, I've got you covered with a 🔥1-minute read summary. ● What is VAR about? It proposes a new autoregressive image generation paradigm. Instead of predicting the next image tokens, it predicts the next image scale. ● What is the evaluation task and metric? Task: ImageNet Image Generation Benchmark. Metric: FID and IS. ● Why is it a milestone work? This is the first time that GPT-style image generation has beaten diffusion models on both image quality and inference speed. ● Why it has huge potential? Just like LLM, this model is proven to follow the scaling law. ● What's next? I can already foresee text-control VAR, video-based VAR, VAR for world models, super super-large VAR coming... For more insights, check out the paper here: https://lnkd.in/gGqdiS9H Modeling and the code source: https://lnkd.in/gTiaMjKG

309