'It's going much too fast': the inside story of the race to create the ultimate AI On the 8.49am train through Silicon Valley, the tables are packed with young people glued to laptops, earbuds in, rattling out code. As the northern California hills scroll past, instructions flash up on screens from bosses: fix this bug; add new script. There is no time to enjoy the view. These commuters are foot soldiers in the global race towards artificial general intelligence - when AI systems become as or more capable than highly qualified humans. Here in the Bay Area of San Francisco, some of the world's biggest companies are fighting it out to gain some kind of an advantage. And, in turn, they are competing with China. This race to seize control of a technology that could reshape the world is being fuelled by bets in the trillions of dollars by the US's most powerful capitalists. Passengers get off a train at Palo Alto station.

The vast majority of deep models use multiple gradient signals, typically corresponding to a sum of multiple loss terms, to update a shared set of trainable weights.

The autonomic nervous system (ANS) is activated during stress, which can have negative effects on cardiovascular health, sleep, the immune system, and mental health. While there are ways to quantify ANS activity in laboratories, there is a paucity of methods that have been validated in real-world contexts. We present the Fitbit Body Response Algorithm, an approach to continuous remote measurement of ANS activation through widely available remote wrist-based sensors. The design was validated via two experiments, a Trier Social Stress Test (n = 45) and ecological momentary assessments (EMA) of perceived stress (n=87), providing both controlled and ecologically valid test data. Model performance predicting perceived stress when using all available sensor modalities was consistent with expectations (accuracy=0.85) and outperformed models with access to only a subset of the signals. We discuss and address challenges to sensing that arise in real world settings that do not present in conventional lab environments.

Waymo is expanding to new (but actually old) territory. The Waymo One service will soon be available in more of the San Francisco Bay Area, specifically Mountain View, Los Altos, Palo Alto and parts of Sunnyvale. The company spent several years testing its self-driving cars in Mountain View, the city where its headquarters is located. According to TechCrunch, Waymo One will be available across 27 square miles of Silicon Valley, in addition to the 55 square miles it covers elsewhere in the Bay Area, including San Francisco. This is the latest in a string of expansions for the company. Waymo has been up and running in Los Angeles and Phoenix for a while.

Evaluation of policies in recommender systems typically involves A/B testing using live experiments on real users to assess a new policy's impact on relevant metrics. This ``gold standard'' comes at a high cost, however, in terms of cycle time, user cost, and potential user retention. In developing policies for ``onboarding'' new users, these costs can be especially problematic, since on-boarding occurs only once. In this work, we describe a simulation methodology used to augment (and reduce) the use of live experiments. We illustrate its deployment for the evaluation of ``preference elicitation'' algorithms used to onboard new users of the YouTube Music platform. By developing counterfactually robust user behavior models, and a simulation service that couples such models with production infrastructure, we are able to test new algorithms in a way that reliably predicts their performance on key metrics when deployed live. We describe our domain, our simulation models and platform, results of experiments and deployment, and suggest future steps needed to further realistic simulation as a powerful complement to live experiments.

Understanding the nuances of a user's extensive interaction history is key to building accurate and personalized natural language systems that can adapt to evolving user preferences. To address this, we introduce PERSOMA, Personalized Soft Prompt Adapter architecture. Unlike previous personalized prompting methods for large language models, PERSOMA offers a novel approach to efficiently capture user history. It achieves this by resampling and compressing interactions as free form text into expressive soft prompt embeddings, building upon recent research utilizing embedding representations as input for LLMs. We rigorously validate our approach by evaluating various adapter architectures, first-stage sampling strategies, parameter-efficient tuning techniques like LoRA, and other personalization methods. Our results demonstrate PERSOMA's superior ability to handle large and complex user histories compared to existing embedding-based and text-prompt-based techniques.

Image generation models are poised to become ubiquitous in a range of applications. These models are often fine-tuned and evaluated using human quality judgments that assume a universal standard, failing to consider the subjectivity of such tasks. To investigate how to quantify subjectivity, and the scale of its impact, we measure how assessments differ among human annotators across different use cases. Simulating the effects of ordinarily latent elements of annotators subjectivity, we contrive a set of motivations (t-shirt graphics, presentation visuals, and phone background images) to contextualize a set of crowdsourcing tasks. Our results show that human evaluations of images vary within individual contexts and across combinations of contexts. Three key factors affecting this subjectivity are image appearance, image alignment with text, and representation of objects mentioned in the text. Our study highlights the importance of taking individual users and contexts into account, both when building and evaluating generative models

This is a hybrid role located at Wayfair's Boston, MA or Mountain View, CA Office. Wayfair's Recommendations team provides the core platforms and services that allow our customers to discover and buy the products they love by serving the right content to every customer on every touchpoint. It also allows our suppliers to surface their products to the right customer to showcase their inventory and build their brand. The Product Recommendations team is searching for a Data Science/Machine learning scientist to join our sort optimization team. This team works on multi-objective optimizations which strike the balance of showing popular products and showcasing sponsored products or brand-aware display ads in the product search and recommendations.

The AI age is upon us and high performance computing is the underlying platform powering everything from Large Language Models (LLM) to Image synthesis from text. However, with the demise of Moore's law and Dennard scaling we are at an inflection point. At Lightmatter, we are leading the transition of computing from traditional electronic transistors to photonic technologies which can operate at mind blowing efficiency and throughput. In this role, you will support all the activities of the ML team as it guides the development of a new class of computing infrastructure. This includes fine tuning LLMs, enablement of new models on custom architectures, evaluating the performance of models at scale, developing abstract models of the hardware for evaluating accuracy and throughput and help co-design novel hardware in a new paradigm of computing.