Articles for category: AI Research

March 11, 2025

ikayaniaamirshahzad@gmail.com

A Multi-Faceted Evaluation for Text-to-Image Diffusion Model Unlearning

[Submitted on 8 Oct 2024 (v1), last revised 9 Mar 2025 (this version, v2)] View a PDF of the paper titled Holistic Unlearning Benchmark: A Multi-Faceted Evaluation for Text-to-Image Diffusion Model Unlearning, by Saemi Moon and 3 other authors View PDF HTML (experimental) Abstract:As text-to-image diffusion models gain widespread commercial applications, there are increasing concerns about unethical or harmful use, including the unauthorized generation of copyrighted or sensitive content. Concept unlearning has emerged as a promising solution to these challenges by removing undesired and harmful information from the pre-trained model. However, the previous evaluations primarily focus on whether target concepts

March 11, 2025

ikayaniaamirshahzad@gmail.com

Google Cloud gen AI technology helps healthcare organizations

Healthcare used to lag behind other industries in tech adoption — but AI is different. Generative AI (gen AI) can improve many of the administrative burdens in healthcare, like finding the right information in a medical record, drafting paperwork or identifying candidates for a clinical trial. And as gen AI continues to evolve from a buzzword into a business essential, new advancements like AI agents, AI-powered search and AI platforms are driving new opportunities for healthcare transformation. At the HIMSS 25 healthcare conference this week, we shared how healthcare organizations are using Google Cloud’s gen AI to build agents and

March 11, 2025

ikayaniaamirshahzad@gmail.com

Reddit – Dive into anything

We value your privacy Reddit and its partners use cookies and similar technologies to provide you with a better experience. By accepting all cookies, you agree to our use of cookies to deliver and maintain our services and site, improve the quality of Reddit, personalize Reddit content and advertising, and measure the effectiveness of advertising. By rejecting non-essential cookies, Reddit may still use certain cookies to ensure the proper functionality of our platform. For more information, please see our Cookie Notice and our Privacy Policy. Source link

March 11, 2025

ikayaniaamirshahzad@gmail.com

Applying GRPO to Qwen-0.5B-Instruct using GSM8K ends up outputting a low-performing model.

For context: I had just read and learned about GRPO last week. This week, I decided to apply this method by training Qwen-0.5B-Instruct on the GSM8K dataset. Using GRPOTrainer from TRL, I set 2 training epochs and reference model synch every 25 steps. I only used two reward functions: strict formatting (i.e., must follow <reasoning>…</reasoning><answer>…</answer> format) and accuracy (i.e., must output the correct answer). However when I tried to ask it a simple question after training phase was done, it wasn't able to answer it. It just instead answers \n (newline) character. I checked the graphs of the reward function

March 10, 2025

ikayaniaamirshahzad@gmail.com

Reddit – Dive into anything

We value your privacy Reddit and its partners use cookies and similar technologies to provide you with a better experience. By accepting all cookies, you agree to our use of cookies to deliver and maintain our services and site, improve the quality of Reddit, personalize Reddit content and advertising, and measure the effectiveness of advertising. By rejecting non-essential cookies, Reddit may still use certain cookies to ensure the proper functionality of our platform. For more information, please see our Cookie Notice and our Privacy Policy. Source link

March 10, 2025

ikayaniaamirshahzad@gmail.com

Reddit – Dive into anything

We value your privacy Reddit and its partners use cookies and similar technologies to provide you with a better experience. By accepting all cookies, you agree to our use of cookies to deliver and maintain our services and site, improve the quality of Reddit, personalize Reddit content and advertising, and measure the effectiveness of advertising. By rejecting non-essential cookies, Reddit may still use certain cookies to ensure the proper functionality of our platform. For more information, please see our Cookie Notice and our Privacy Policy. Source link

March 10, 2025

ikayaniaamirshahzad@gmail.com

Reddit – Dive into anything

We value your privacy Reddit and its partners use cookies and similar technologies to provide you with a better experience. By accepting all cookies, you agree to our use of cookies to deliver and maintain our services and site, improve the quality of Reddit, personalize Reddit content and advertising, and measure the effectiveness of advertising. By rejecting non-essential cookies, Reddit may still use certain cookies to ensure the proper functionality of our platform. For more information, please see our Cookie Notice and our Privacy Policy. Source link

March 10, 2025

ikayaniaamirshahzad@gmail.com

Reddit – Dive into anything

We value your privacy Reddit and its partners use cookies and similar technologies to provide you with a better experience. By accepting all cookies, you agree to our use of cookies to deliver and maintain our services and site, improve the quality of Reddit, personalize Reddit content and advertising, and measure the effectiveness of advertising. By rejecting non-essential cookies, Reddit may still use certain cookies to ensure the proper functionality of our platform. For more information, please see our Cookie Notice and our Privacy Policy. Source link

March 10, 2025

ikayaniaamirshahzad@gmail.com

HEAL: A framework for health equity assessment of machine learning performance

Posted by Mike Schaekermann, Research Scientist, Google Research, and Ivor Horn, Chief Health Equity Officer & Director, Google Core Health equity is a major societal concern worldwide with disparities having many causes. These sources include limitations in access to healthcare, differences in clinical treatment, and even fundamental differences in the diagnostic technology. In dermatology for example, skin cancer outcomes are worse for populations such as minorities, those with lower socioeconomic status, or individuals with limited healthcare access. While there is great promise in recent advances in machine learning (ML) and artificial intelligence (AI) to help improve healthcare, this transition from

March 10, 2025

ikayaniaamirshahzad@gmail.com

Semantic Telemetry: Understanding how users interact with AI systems

AI tools are proving useful across a range of applications, from helping to drive the new era of business transformation to helping artists craft songs. But which applications are providing the most value to users? We’ll dig into that question in a series of blog posts that introduce the Semantic Telemetry project at Microsoft Research. In this initial post, we will introduce a new data science approach that we will use to analyze topics and task complexity of Copilot in Bing usage. Human-AI interactions can be iterative and complex, requiring a new data science approach to understand user behavior to