General-purpose robots need a versatile body and an intelligent mind. Recent advancements in humanoid robots have shown great promise as a hardware platform for building generalist autonomy in the human world. A robot foundation model, trained on massive and diverse data sources, is essential for enabling the robots to reason about novel situations, robustly handle real-world variability, and rapidly learn new tasks. To this end, we introduce GR00T N1, an open foundation model for humanoid robots. GR00T N1 is a Vision-Language-Action (VLA) model with a dual-system architecture. The vision-language module (System 2) interprets the environment through vision and language instructions. The subsequent diffusion transformer module (System 1) generates fluid motor actions in real time. Both modules are tightly coupled and jointly trained end-to-end. We train GR00T N1 with a heterogeneous mixture of real-robot trajectories, human videos, and synthetically generated datasets. We show that our generalist robot model GR00T N1 outperforms the state-of-the-art imitation learning baselines on standard simulation benchmarks across multiple robot embodiments. Furthermore, we deploy our model on the Fourier GR-1 humanoid robot for language-conditioned bimanual manipulation tasks, achieving strong performance with high data efficiency.

State-of-the-art LLMs are powered by scaling -- scaling model size, dataset size and cluster size. It is economically infeasible to extensively tune hyperparameter for the largest runs. Instead, approximately optimal hyperparameters must be inferred or \textit{transferred} from smaller experiments. Hyperparameter transfer across model sizes has been studied in Yang et al. However, hyperparameter transfer across dataset size -- or token horizon -- has not been studied yet. To remedy this we conduct a large scale empirical study on how optimal learning rate (LR) depends on token horizon in LLM training. We first demonstrate that the optimal LR changes significantly with token horizon -- longer training necessitates smaller LR. Secondly we demonstrate the the optimal LR follows a scaling law, and that the optimal LR for longer horizons can be accurately estimated from shorter horizons via such scaling laws. We also provide a rule-of-thumb for transferring LR across token horizons with zero overhead over current practices. Lastly we provide evidence that LLama-1 used too high LR, and estimate the performance hit from this. We thus argue that hyperparameter transfer across data size is an important and overlooked component of LLM training.

We introduce phi-3-mini, a 3.8 billion parameter language model trained on 3.3 trillion tokens, whose overall performance, as measured by both academic benchmarks and internal testing, rivals that of models such as Mixtral 8x7B and GPT-3.5 (e.g., phi-3-mini achieves 69% on MMLU and 8.38 on MT-bench), despite being small enough to be deployed on a phone. The innovation lies entirely in our dataset for training, a scaled-up version of the one used for phi-2, composed of heavily filtered publicly available web data and synthetic data. The model is also further aligned for robustness, safety, and chat format. We also provide some initial parameter-scaling results with a 7B and 14B models trained for 4.8T tokens, called phi-3-small and phi-3-medium, both significantly more capable than phi-3-mini (e.g., respectively 75% and 78% on MMLU, and 8.7 and 8.9 on MT-bench). Moreover, we also introduce phi-3-vision, a 4.2 billion parameter model based on phi-3-mini with strong reasoning capabilities for image and text prompts.

A big convergence of language, multimodal perception, action, and world modeling is a key step toward artificial general intelligence. In this work, we introduce Kosmos-1, a Multimodal Large Language Model (MLLM) that can perceive general modalities, learn in context (i.e., few-shot), and follow instructions (i.e., zero-shot). Specifically, we train Kosmos-1 from scratch on web-scale multimodal corpora, including arbitrarily interleaved text and images, image-caption pairs, and text data. We evaluate various settings, including zero-shot, few-shot, and multimodal chain-of-thought prompting, on a wide range of tasks without any gradient updates or finetuning. Experimental results show that Kosmos-1 achieves impressive performance on (i) language understanding, generation, and even OCR-free NLP (directly fed with document images), (ii) perception-language tasks, including multimodal dialogue, image captioning, visual question answering, and (iii) vision tasks, such as image recognition with descriptions (specifying classification via text instructions). We also show that MLLMs can benefit from cross-modal transfer, i.e., transfer knowledge from language to multimodal, and from multimodal to language. In addition, we introduce a dataset of Raven IQ test, which diagnoses the nonverbal reasoning capability of MLLMs.

Reinforcement learning (RL) experiments have notoriously high variance, and minor details can have disproportionately large effects on measured outcomes. This is problematic for creating reproducible research and also serves as an obstacle for real-world applications, where safety and predictability are paramount. In this paper, we investigate causes for this perceived instability. To allow for an in-depth analysis, we focus on a specifically popular setup with high variance -- continuous control from pixels with an actor-critic agent. In this setting, we demonstrate that variance mostly arises early in training as a result of poor "outlier" runs, but that weight initialization and initial exploration are not to blame. We show that one cause for early variance is numerical instability which leads to saturating nonlinearities. We investigate several fixes to this issue and find that one particular method is surprisingly effective and simple -- normalizing penultimate features. Addressing the learning instability allows for larger learning rates, and significantly decreases the variance of outcomes. This demonstrates that the perceived variance in RL is not necessarily inherent to the problem definition and may be addressed through simple architectural modifications.

Batch normalization is a ubiquitous deep learning technique that normalizes activations in intermediate layers. It is associated with improved accuracy and faster learning, but despite its enormous success there is little consensus regarding why it works. We aim to rectify this and take an empirical approach to understanding batch normalization. Our primary observation is that the higher learning rates that batch normalization enables have a regularizing effect that dramatically improves generalization of normalized networks, which is both demonstrated empirically and motivated theoretically. We show how activations become large and how the convolutional channels become increasingly ill-behaved for layers deep in unnormalized networks, and how this results in larger input-independent gradients. Beyond just gradient scaling, we demonstrate how the learning rate in unnormalized networks is further limited by the magnitude of activations growing exponentially with network depth for large parameter updates, a problem batch normalization trivially avoids. Motivated by recent results in random matrix theory, we argue that ill-conditioning of the activations is due to fluctuations in random initialization, shedding new light on classical initialization schemes and their consequences.

From the stone age, to the bronze, iron age, and modern silicon age, the discovery and characterization of new materials has always been instrumental to humanity's progress and development. With the current pressing need to address sustainability challenges and find alternatives to fossil fuels, we look for solutions in the development of new materials that will allow for renewable energy. To discover materials with the required properties, materials scientists can perform high-throughput materials discovery, which includes rapid synthesis and characterization via X-ray diffraction (XRD) of thousands of materials. A central problem in materials discovery, the phase map identification problem, involves the determination of the crystal structure of materials from materials composition and structural characterization data. This analysis is traditionally performed mainly by hand, which can take days for a single material system. In this work we present Phase-Mapper, a solution platform that tightly integrates XRD experimentation, AI problem solving, and human intelligence for interpreting XRD patterns and inferring the crystal structures of the underlying materials. Phase-Mapper is compatible with any spectral demixing algorithm, including our novel solver, AgileFD, which is based on convolutive non-negative matrix factorization. AgileFD allows materials scientists to rapidly interpret XRD patterns, and incorporates constraints to capture prior knowledge about the physics of the materials as well as human feedback. With our system, materials scientists have been able to interpret previously unsolvable systems of XRD data at the Department of Energy’s Joint Center for Artificial Photosynthesis, including the Nb-Mn-V oxide system, which led to the discovery of new solar light absorbers and is provided as an illustrative example of AI-enabled high throughput materials discovery

Cascades represent rapid changes in networks. A cascading phenomenon of ecological and economic impact is the spread of invasive species in geographic landscapes. The most promising management strategy is often biocontrol, which entails introducing a natural predator able to control the invading population, a setting that can be treated as two interacting cascades of predator and prey populations. We formulate and study a nonlinear problem of optimal biocontrol: optimally seeding the predator cascade over time to minimize the harmful prey population. Recurring budgets, which typically face conservation organizations, naturally leads to sparse constraints which make the problem amenable to approximation algorithms. Available methods based on continuous relaxations scale poorly, to remedy this we develop a novel and scalable randomized algorithm based on a width relaxation, applicable to a broad class of combinatorial optimization problems. We evaluate our contributions in the context of biocontrol for the insect pest Hemlock Wolly Adelgid (HWA) in eastern North America. Our algorithm outperforms competing methods in terms of scalability and solution quality and finds near-optimal strategies for the control of the HWA for fine-grained networks -- an important problem in computational sustainability.

High-throughput materials discovery involves the rapid synthesis, measurement, and characterization of many different but structurally related materials. A central problem in materials discovery, the phase map identification problem, involves the determination of the crystal structure of materials from materials composition and structural characterization data. We present Phase-Mapper, a novel solution platform that allows humans to interact with both the data and products of AI algorithms, including the incorporation of human feedback to constrain or initialize solutions. Phase-Mapper is compatible with any spectral demixing algorithm, including our novel solver, AgileFD, which is based on convolutive non-negative matrix factorization. AgileFD allows materials scientists to rapidly interpret XRD patterns, and can incorporate constraints to capture the physics of the materials as well as human feedback. We compare three solver variants with previously proposed methods in a large-scale experiment involving 20 synthetic systems, demonstrating the efficacy of imposing physical constraints using AgileFD. Since the deployment of Phase-Mapper at the Department of Energy's Joint Center for Artificial Photosynthesis (JCAP), thousands of X-ray diffraction patterns have been processed and the results are yielding discovery of new materials for energy applications, as exemplified by the discovery of a new family of metal oxide solar light absorbers, among the previously unsolved Nb-Mn-V oxide system, which is provided here as an illustrative example. Phase-Mapper is also being deployed at the Stanford Synchrotron Radiation Lightsource (SSRL) to enable phase mapping on datasets in real time.