This paper was accepted at the Workshop on Foundation Models in the Wild at ICLR 2025. Visual understanding is inherently contextual – what we focus on in an image depends on the task at hand. For instance, given an image of a person holding a bouquet of flowers, we may focus on either the person …
Read more “FocalLens: Instruction Tuning Enables Zero-Shot Conditional Image Representations”
This post is co-written with Kim Nguyen and Shyam Banuprakash from Clario. Clario is a leading provider of endpoint data solutions to the clinical trials industry, generating high-quality clinical evidence for life sciences companies seeking to bring new therapies to patients. Since Clario’s founding more than 50 years ago, the company’s endpoint data solutions have …
Not all AI scaling strategies are equal. Longer reasoning chains are not sign of higher intelligence. More compute isn’t always the answer.Read More
A lawsuit over the Trump administration’s infamous Houthi Signal group chat has revealed what steps departments took to preserve the messages—and how little they actually saved.
A team has developed an explainable AI model for automatic collision avoidance between ships.
Penn Engineers have developed the first programmable chip that can train nonlinear neural networks using light—a breakthrough that could dramatically speed up AI training, reduce energy use and even pave the way for fully light-powered computers.
This post is divided into three parts; they are: • Understanding Context Vectors • Visualizing Context Vectors from Different Layers • Visualizing Attention Patterns Unlike traditional word embeddings (such as Word2Vec or GloVe), which assign a fixed vector to each word regardless of context, transformer models generate dynamic representations that depend on surrounding words.
Retrieval augmented generation (RAG) is one of 2025’s hot topics in the AI landscape.
Be sure to check out the previous articles in this series: •
At Apple, we believe privacy is a fundamental human right. And we believe in giving our users a great experience while protecting their privacy. For years, we’ve used techniques like differential privacy as part of our opt-in device analytics program. This lets us gain insights into how our products are used, so we can improve …
Read more “Understanding Aggregate Trends for Apple Intelligence Using Differential Privacy”