Halim A.

Interesting paper about the limitations of the generative AI. Actually Yann LeCun talks about this limitation as inherited problem in the existing models for long time. Knowing the limitation helps to align expectations. https://lnkd.in/dCid9B_J #generativeai #llm #artificailintelligence #ai

7

3 Comments

🚀 We are thrilled to announce the release of our new paper: "The Evolution of Multimodal Model Architectures." In this paper, we provide a bird's-eye view of all the state-of-the-art multimodal (large) models that have emerged since the inception of Transformers. It highlights the pros and cons of each model architecture, grouping them into four broad categories. These categories can serve as a blueprint to help you determine which architecture is best suited for your specific scenario. A huge thank you to all the co-authors and especially to Shakti Wadekar for putting this paper together during his PhD at Purdue University. Paper: 🔗 [https://lnkd.in/gZHQUS6c] Authors: Shakti Wadekar, Abhishek Chaurasia, Aman Chadha, and Eugenio Culurciello #AI #MachineLearning #MultimodalModels #Transformers #Research #PurdueUniversity

80

4 Comments

This approach could turn the Generative AI industry upside down. Most everyone knows that AI requires computers to do a lot of calculations and especially for Large Language Models (LLMs) like ChatGPT. The majority of these calculations are actually “Matrix Multiplications”. (You remember, a 3x3 matrix times another 3x3 matrix). If the researchers in the article are proven correct, they have demonstrated a way that Matrix Multiplication can be replaced with a more efficient alternative inside the LLM getting nearly the same results. The article positions this savings as a win for the environment (yeah environment), but I think they are missing the boat. More efficient means less electricity and potentially a smaller need for GPU resources, the bread and butter of stock darling Nvidia. When you combine this cost saving with reduced privacy concerns of hosted LLMs (Open AI) and innovative LLM licensing schemes (LibreChat - https://lnkd.in/gKQjMfz5) you can see the cracks in the current direction of the LLM industry. Q: If you could run your own LLM on your own private environment with MIT open-source licensing and without expensive GPUs, why exactly do you need Nvidia, Open AI and alike? Bold prediction: if the researchers get this working at scale, we could see a rapid shift and commoditization in the LLM and generative AI space. All the pieces are in place. TLDR details from the article, describing the innovation: “The researchers' approach involves two main innovations: first, they created a custom LLM and constrained it to use only ternary values (-1, 0, 1) instead of traditional floating-point numbers, which allows for simpler computations. Second, the researchers redesigned the computationally expensive self-attention mechanism in traditional language models with a simpler, more efficient unit (that they called a MatMul-free Linear Gated Recurrent Unit—or MLGRU) that processes words sequentially using basic arithmetic operations instead of matrix multiplications.” #AI #GenerativeAI #Innovation

24

Proud today to announce Concentriq Embeddings: Proscia's Concentriq platform now enables generation of foundation model embeddings from pathology images right where your data lives. No learning how to use a bunch of new tools, fumbling with massive infrastructure and compatibility issues every time you want to try out a new model, or wasting precious hours waiting on downloads--a collection of foundation models is available at your fingertips. This launch marks a monumental first step in our vision to make AI for pathology effortless - from proof of concept to development to deployment. Your AI work should happen where the rest of your work does. The future is as bold as the difference between sending a telegraph and sharing video with a smartphone. We can't wait to see what our users will build. 🚀 Congrats to the team that has grown this idea into an epic reality! Special kudos to Kyriakos Toulgaridis Corey Chivers Vaughn Spurrier Jeff Baatz Casey Wahl Nate Langlois Ashley Faber Elbert Basolis III

20

2 Comments

I'll do even a braver poll than last, but only because of this Eric Topol, MD article in Science Magazine :) Multimodal large AI models will be used in medical forecasting (disease risk, onset, trajectory) A) now B) soon but at population level only C) sooner than in diagnostics D) later than in diagnostics E) never https://lnkd.in/gGBHSptv

7

👉🏼 How effective are large language models in product risk assessment? 🤓 Zachary A Collier 👇🏻 https://lnkd.in/eZy9nVrh 🔍 Focus on data insights: - Large language models, like ChatGPT, show promise in brainstorming potential failure modes and risk mitigations. - However, there were errors and inconsistencies in some results. - The guidance provided by the models was perceived as overly generic and occasionally outlandish. 💡 Main outcomes and implications: - ChatGPT performed better at divergent thinking tasks but lacked depth of knowledge compared to subject matter experts. - Similar patterns in strengths and weaknesses were observed when comparing ChatGPT to other large language models. - Despite challenges, there may be a role for large language models in assisting with ideation in product risk assessment. 📚 Field significance: - Product safety professionals can leverage large language models for brainstorming and idea generation in risk assessment processes. - Experts should focus on critically reviewing AI-generated content to ensure accuracy and relevance. 🗄️: [#productrisks #AI #riskassessment #ChatGPT]

1

1 Comment

Responsible and accountable AI models are essential as generative AI continues to revolutionize industries. The rise of this technology, however, presents challenges in understanding how these models make decisions. Mechanistic interpretability addresses this by going beyond traditional methods that track statistical relationships. A recent research work by a team from #Anthropic has involved visualizing which neurons activate in response to specific prompts, advancing monosemanticity—ensuring each neuron’s activation correlates to a single, clear function. This step is crucial for fostering more transparent and safer AI systems in various applications. This is a great progress toward interpreting the thinking mechanisms of large language models (LLMs). #GenerativeAI #AI #LLMs #GenerativeAI #Grainger #MachineLearning #DeepLearning #ExplainableAI #DataScience

45

2 Comments

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

Organizations have insufficient technical capabilities for the speed and scale of their ML model implementations. My colleague Sumit Agarwal just published a very helpful research with a capability view and sample vendors for defining a modular, flexible and scalable MLOps solution. https://lnkd.in/gnB_T9bq #MLOps #LLMOps #AI

149

17 Comments

Interesting that on the MIMIC IV readmission models, the LLMs get closer to the traditional ML models. Perhaps because LLMs are not great at representing numeric values (especially at the model sizes tested) that are informative of the prediction for acute outcomes, whereas readmission relies more on diagnoses and other more linguistically-coded inputs. (Though that doesn't explain the poor performance on MIMIC-III readmission)

3

If you think compute scaling will plateau, think again. We are in the baby states of inference time compute. VLLMs coupled with computer tools, reward learning and RL can leverage orders of magnitude more inference time compute than what we have seen with foundation models. Here’s a concrete training architecture (replace \pi with a large pre-trained vLLM and add it to the reward learner too, \R is the space of all computer tools)

21

1 Comment

"Why did the neural network go to therapy? It had too many layers of unresolved issues! 😂 On a more serious note, whether you’re dealing with deep learning models or business challenges, it's all about peeling back those layers to find clarity and solutions. #AI #MachineLearning #NeuralNetworks #TechHumor #DataScience"

12

simple, and useful. Help you extract table data from Arxiv paper.

2

You had us at bringing up the infinite regress of programs testing programs that test programs! Check out how one can ameliorate that paradox of AI safety using logic and algebra as shown in the ntqr Python package documentation page (https://lnkd.in/e_9fs_Js)