D. Sculley

Sandhini Agarwal, Lama Ahmad, Jason Ai et al. · openai

We present gpt-oss-120b and gpt-oss-20b, two open-weight reasoning models that push the frontier of accuracy and inference cost. The models use an efficient mixture-of-expert transformer architecture and are trained using large-scale distillation and reinforcement learning. We optimize the models to have strong agentic capabilities (deep research browsing, python tool use, and support for developer-provided functions), all while using a rendered chat format that enables clear instruction following and role delineation. Both models achieve strong results on benchmarks ranging from mathematics, coding, and safety. We release the model weights, inference implementations, tool environments, and tokenizers under an Apache 2.0 license to enable broad use and further research.

Gemma Team, Morgane Riviere, Shreya Pathak et al. · deepmind

In this work, we introduce Gemma 2, a new addition to the Gemma family of lightweight, state-of-the-art open models, ranging in scale from 2 billion to 27 billion parameters. In this new version, we apply several known technical modifications to the Transformer architecture, such as interleaving local-global attentions (Beltagy et al., 2020a) and group-query attention (Ainslie et al., 2023). We also train the 2B and 9B models with knowledge distillation (Hinton et al., 2015) instead of next token prediction. The resulting models deliver the best performance for their size, and even offer competitive alternatives to models that are 2-3 times bigger. We release all our models to the community.

Dustin Tran, Jeremiah Liu, Michael W. Dusenberry et al. · oxford

A recent trend in artificial intelligence is the use of pretrained models for language and vision tasks, which have achieved extraordinary performance but also puzzling failures. Probing these models' abilities in diverse ways is therefore critical to the field. In this paper, we explore the reliability of models, where we define a reliable model as one that not only achieves strong predictive performance but also performs well consistently over many decision-making tasks involving uncertainty (e.g., selective prediction, open set recognition), robust generalization (e.g., accuracy and proper scoring rules such as log-likelihood on in- and out-of-distribution datasets), and adaptation (e.g., active learning, few-shot uncertainty). We devise 10 types of tasks over 40 datasets in order to evaluate different aspects of reliability on both vision and language domains. To improve reliability, we developed ViT-Plex and T5-Plex, pretrained large model extensions for vision and language modalities, respectively. Plex greatly improves the state-of-the-art across reliability tasks, and simplifies the traditional protocol as it improves the out-of-the-box performance and does not require designing scores or tuning the model for each task. We demonstrate scaling effects over model sizes up to 1B parameters and pretraining dataset sizes up to 4B examples. We also demonstrate Plex's capabilities on challenging tasks including zero-shot open set recognition, active learning, and uncertainty in conversational language understanding.

Gemma Team, Aishwarya Kamath, Johan Ferret et al. · deepmind, mit

We introduce Gemma 3, a multimodal addition to the Gemma family of lightweight open models, ranging in scale from 1 to 27 billion parameters. This version introduces vision understanding abilities, a wider coverage of languages and longer context - at least 128K tokens. We also change the architecture of the model to reduce the KV-cache memory that tends to explode with long context. This is achieved by increasing the ratio of local to global attention layers, and keeping the span on local attention short. The Gemma 3 models are trained with distillation and achieve superior performance to Gemma 2 for both pre-trained and instruction finetuned versions. In particular, our novel post-training recipe significantly improves the math, chat, instruction-following and multilingual abilities, making Gemma3-4B-IT competitive with Gemma2-27B-IT and Gemma3-27B-IT comparable to Gemini-1.5-Pro across benchmarks. We release all our models to the community.

Zachary Nado, Neil Band, Mark Collier et al.

High-quality estimates of uncertainty and robustness are crucial for numerous real-world applications, especially for deep learning which underlies many deployed ML systems. The ability to compare techniques for improving these estimates is therefore very important for research and practice alike. Yet, competitive comparisons of methods are often lacking due to a range of reasons, including: compute availability for extensive tuning, incorporation of sufficiently many baselines, and concrete documentation for reproducibility. In this paper we introduce Uncertainty Baselines: high-quality implementations of standard and state-of-the-art deep learning methods on a variety of tasks. As of this writing, the collection spans 19 methods across 9 tasks, each with at least 5 metrics. Each baseline is a self-contained experiment pipeline with easily reusable and extendable components. Our goal is to provide immediate starting points for experimentation with new methods or applications. Additionally we provide model checkpoints, experiment outputs as Python notebooks, and leaderboards for comparing results. Code available at https://github.com/google/uncertainty-baselines.

Shreya Shankar, Yoni Halpern, Eric Breck et al.

Modern machine learning systems such as image classifiers rely heavily on large scale data sets for training. Such data sets are costly to create, thus in practice a small number of freely available, open source data sets are widely used. We suggest that examining the geo-diversity of open data sets is critical before adopting a data set for use cases in the developing world. We analyze two large, publicly available image data sets to assess geo-diversity and find that these data sets appear to exhibit an observable amerocentric and eurocentric representation bias. Further, we analyze classifiers trained on these data sets to assess the impact of these training distributions and find strong differences in the relative performance on images from different locales. These results emphasize the need to ensure geo-representation when constructing data sets for use in the developing world.

Daniel Smilkov, Shan Carter, D. Sculley et al.

The recent successes of deep learning have led to a wave of interest from non-experts. Gaining an understanding of this technology, however, is difficult. While the theory is important, it is also helpful for novices to develop an intuitive feel for the effect of different hyperparameters and structural variations. We describe TensorFlow Playground, an interactive, open sourced visualization that allows users to experiment via direct manipulation rather than coding, enabling them to quickly build an intuition about neural nets.

D. Sculley, Will Cukierski, Phil Culliton et al.

In this position paper, we observe that empirical evaluation in Generative AI is at a crisis point since traditional ML evaluation and benchmarking strategies are insufficient to meet the needs of evaluating modern GenAI models and systems. There are many reasons for this, including the fact that these models typically have nearly unbounded input and output spaces, typically do not have a well defined ground truth target, and typically exhibit strong feedback loops and prediction dependence based on context of previous model outputs. On top of these critical issues, we argue that the problems of leakage and contamination are in fact the most important and difficult issues to address for GenAI evaluations. Interestingly, the field of AI Competitions has developed effective measures and practices to combat leakage for the purpose of counteracting cheating by bad actors within a competition setting. This makes AI Competitions an especially valuable (but underutilized) resource. Now is time for the field to view AI Competitions as the gold standard for empirical rigor in GenAI evaluation, and to harness and harvest their results with according value.

Alicia Parrish, Hannah Rose Kirk, Jessica Quaye et al.

The generative AI revolution in recent years has been spurred by an expansion in compute power and data quantity, which together enable extensive pre-training of powerful text-to-image (T2I) models. With their greater capabilities to generate realistic and creative content, these T2I models like DALL-E, MidJourney, Imagen or Stable Diffusion are reaching ever wider audiences. Any unsafe behaviors inherited from pretraining on uncurated internet-scraped datasets thus have the potential to cause wide-reaching harm, for example, through generated images which are violent, sexually explicit, or contain biased and derogatory stereotypes. Despite this risk of harm, we lack systematic and structured evaluation datasets to scrutinize model behavior, especially adversarial attacks that bypass existing safety filters. A typical bottleneck in safety evaluation is achieving a wide coverage of different types of challenging examples in the evaluation set, i.e., identifying 'unknown unknowns' or long-tail problems. To address this need, we introduce the Adversarial Nibbler challenge. The goal of this challenge is to crowdsource a diverse set of failure modes and reward challenge participants for successfully finding safety vulnerabilities in current state-of-the-art T2I models. Ultimately, we aim to provide greater awareness of these issues and assist developers in improving the future safety and reliability of generative AI models. Adversarial Nibbler is a data-centric challenge, part of the DataPerf challenge suite, organized and supported by Kaggle and MLCommons.

Alexander D'Amour, Katherine Heller, Dan Moldovan et al.

ML models often exhibit unexpectedly poor behavior when they are deployed in real-world domains. We identify underspecification as a key reason for these failures. An ML pipeline is underspecified when it can return many predictors with equivalently strong held-out performance in the training domain. Underspecification is common in modern ML pipelines, such as those based on deep learning. Predictors returned by underspecified pipelines are often treated as equivalent based on their training domain performance, but we show here that such predictors can behave very differently in deployment domains. This ambiguity can lead to instability and poor model behavior in practice, and is a distinct failure mode from previously identified issues arising from structural mismatch between training and deployment domains. We show that this problem appears in a wide variety of practical ML pipelines, using examples from computer vision, medical imaging, natural language processing, clinical risk prediction based on electronic health records, and medical genomics. Our results show the need to explicitly account for underspecification in modeling pipelines that are intended for real-world deployment in any domain.

Zachary Nado, Shreyas Padhy, D. Sculley et al.

Covariate shift has been shown to sharply degrade both predictive accuracy and the calibration of uncertainty estimates for deep learning models. This is worrying, because covariate shift is prevalent in a wide range of real world deployment settings. However, in this paper, we note that frequently there exists the potential to access small unlabeled batches of the shifted data just before prediction time. This interesting observation enables a simple but surprisingly effective method which we call prediction-time batch normalization, which significantly improves model accuracy and calibration under covariate shift. Using this one line code change, we achieve state-of-the-art on recent covariate shift benchmarks and an mCE of 60.28\% on the challenging ImageNet-C dataset; to our knowledge, this is the best result for any model that does not incorporate additional data augmentation or modification of the training pipeline. We show that prediction-time batch normalization provides complementary benefits to existing state-of-the-art approaches for improving robustness (e.g. deep ensembles) and combining the two further improves performance. Our findings are supported by detailed measurements of the effect of this strategy on model behavior across rigorous ablations on various dataset modalities. However, the method has mixed results when used alongside pre-training, and does not seem to perform as well under more natural types of dataset shift, and is therefore worthy of additional study. We include links to the data in our figures to improve reproducibility, including a Python notebooks that can be run to easily modify our analysis at https://colab.research.google.com/drive/11N0wDZnMQQuLrRwRoumDCrhSaIhkqjof.

Daniel Smilkov, Nikhil Thorat, Yannick Assogba et al.

TensorFlow.js is a library for building and executing machine learning algorithms in JavaScript. TensorFlow.js models run in a web browser and in the Node.js environment. The library is part of the TensorFlow ecosystem, providing a set of APIs that are compatible with those in Python, allowing models to be ported between the Python and JavaScript ecosystems. TensorFlow.js has empowered a new set of developers from the extensive JavaScript community to build and deploy machine learning models and enabled new classes of on-device computation. This paper describes the design, API, and implementation of TensorFlow.js, and highlights some of the impactful use cases.

Alexey A. Gritsenko, Alex D'Amour, James Atwood et al.

We introduce the BriarPatch, a pixel-space intervention that obscures sensitive attributes from representations encoded in pre-trained classifiers. The patches encourage internal model representations not to encode sensitive information, which has the effect of pushing downstream predictors towards exhibiting demographic parity with respect to the sensitive information. The net result is that these BriarPatches provide an intervention mechanism available at user level, and complements prior research on fair representations that were previously only applicable by model developers and ML experts.

Jennifer N. Wei, David Belanger, Ryan P. Adams et al.

When confronted with a substance of unknown identity, researchers often perform mass spectrometry on the sample and compare the observed spectrum to a library of previously-collected spectra to identify the molecule. While popular, this approach will fail to identify molecules that are not in the existing library. In response, we propose to improve the library's coverage by augmenting it with synthetic spectra that are predicted using machine learning. We contribute a lightweight neural network model that quickly predicts mass spectra for small molecules. Achieving high accuracy predictions requires a novel neural network architecture that is designed to capture typical fragmentation patterns from electron ionization. We analyze the effects of our modeling innovations on library matching performance and compare our models to prior machine learning-based work on spectrum prediction.

Brian Patton, Yannis Agiomyrgiannakis, Michael Terry et al.

Developers of text-to-speech synthesizers (TTS) often make use of human raters to assess the quality of synthesized speech. We demonstrate that we can model human raters' mean opinion scores (MOS) of synthesized speech using a deep recurrent neural network whose inputs consist solely of a raw waveform. Our best models provide utterance-level estimates of MOS only moderately inferior to sampled human ratings, as shown by Pearson and Spearman correlations. When multiple utterances are scored and averaged, a scenario common in synthesizer quality assessment, AutoMOS achieves correlations approaching those of human raters. The AutoMOS model has a number of applications, such as the ability to explore the parameter space of a speech synthesizer without requiring a human-in-the-loop.

Daniel Golovin, D. Sculley, H. Brendan McMahan et al.

We reduce the memory footprint of popular large-scale online learning methods by projecting our weight vector onto a coarse discrete set using randomized rounding. Compared to standard 32-bit float encodings, this reduces RAM usage by more than 50% during training and by up to 95% when making predictions from a fixed model, with almost no loss in accuracy. We also show that randomized counting can be used to implement per-coordinate learning rates, improving model quality with little additional RAM. We prove these memory-saving methods achieve regret guarantees similar to their exact variants. Empirical evaluation confirms excellent performance, dominating standard approaches across memory versus accuracy tradeoffs.