AI is undergoing a paradigm shift with the rise of models (e.g., BERT, DALL-E, GPT-3) that are trained on broad data at scale and are adaptable to a wide range of downstream tasks. We call these models foundation models to underscore their critically central yet incomplete character. This report provides a thorough account of the opportunities and risks of foundation models, ranging from their capabilities (e.g., language, vision, robotics, reasoning, human interaction) and technical principles (e.g., model architectures, training procedures, data, systems, security, evaluation, theory) to their applications (e.g., law, healthcare, education) and societal impact (e.g., inequity, misuse, economic and environmental impact, legal and ethical considerations). Though foundation models are based on standard deep learning and transfer learning, their scale results in new emergent capabilities, and their effectiveness across so many tasks incentivizes homogenization. Homogenization provides powerful leverage but demands caution, as the defects of the foundation model are inherited by all the adapted models downstream. Despite the impending widespread deployment of foundation models, we currently lack a clear understanding of how they work, when they fail, and what they are even capable of due to their emergent properties. To tackle these questions, we believe much of the critical research on foundation models will require deep interdisciplinary collaboration commensurate with their fundamentally sociotechnical nature.

We report on the creation of a dataset for studying assessment of suicide risk via online postings in Reddit. Evaluation of risk-level annotations by experts yields what is, to our knowledge, the first demonstration of reliability in risk assessment by clinicians based on social media postings. We also introduce and demonstrate the value of a new, detailed rubric for assessing suicide risk, compare crowdsourced with expert performance, and present baseline predictive modeling experiments using the new dataset, which will be made available to researchers through the American Association of Suicidology.

In this work, we propose a novel robot learning framework called Neural Task Programming (NTP), which bridges the idea of few-shot learning from demonstration and neural program induction. NTP takes as input a task specification (e.g., video demonstration of a task) and recursively decomposes it into finer sub-task specifications. These specifications are fed to a hierarchical neural program, where bottomlevel programs are callable subroutines that interact with the environment. We validate our method in three robot manipulation tasks. NTP achieves strong generalization across sequential tasks that exhibit hierarchal and compositional structures. The experimental results show that NTP learns to generalize well towards unseen tasks with increasing lengths, variable topologies, and changing objectives. stanfordvl.github.io/ntp/ arXiv:1710.01813v2 [cs.AI]

Our goal is to generate a policy to complete an unseen task given just a single video demonstration of the task in a given domain. We hypothesize that to successfully generalize to unseen complex tasks from a single video demonstration, it is necessary to explicitly incorporate the compositional structure of the tasks into the model. To this end, we propose Neural Task Graph (NTG) Networks, which use conjugate task graph as the intermediate representation to modularize both the video demonstration and the derived policy. We empirically show NTG achieves inter-task generalization on two complex tasks: Block Stacking in Bul-letPhysics and Object Collection in AI2-THOR. NTG improves data efficiency with visual input as well as achieve strong generalization without the need for dense hierarchical supervision. We further show that similar performance trends hold when applied to real-world data. We show that NTG can effectively predict task structure on the JIGSAWS surgical dataset and generalize to unseen tasks.

In earthquake early warning (EEW), every sufficiently impulsive signal is potentially the first evidence for an unfolding large earthquake. More often than not, however, impulsive signals are mere nuisance signals. One of the most fundamental-and difficult-tasks in EEW is to rapidly and reliably discriminate real local earthquake signals from all other signals. This discrimination is necessarily based on very little information, typically a few seconds worth of seismic waveforms from a small number of stations. As a result, current EEW systems struggle to avoid discrimination errors and suffer from false and missed alerts. In this study we show how modern machine learning classifiers can strongly improve real-time signal/noise discrimination. We develop and compare a series of nonlinear classifiers with variable architecture depths, including fully connected, convolutional and recurrent neural networks, and a model that combines a generative adversarial network with a random forest. We train all classifiers on the same data set, which includes 374 k local earthquake records (M3.0-9.1) and 946 k impulsive noise signals. We find that all classifiers outperform existing simple linear classifiers and that complex models trained directly on the raw signals yield the greatest degree of improvement. Using 3-s-long waveform snippets, the convolutional neural network and the generative adversarial network with a random forest classifiers both reach 99.5% precision and 99.3% recall on an independent validation data set. Most misclassifications stem from impulsive teleseismic records, and from incorrectly labeled records in the data set. Our results suggest that machine learning classifiers can strongly improve the reliability and speed of EEW alerts.Plain Language Summary Seismic stations record not only earthquake signals but also a wide variety of nuisance signals. Some of these nuisance signals are impulsive and can initially look very similar to real earthquake signals. This is a problem for earthquake early warning (EEW) algorithms, which sometimes misinterpret such signals as being real earthquake signals, and which may then send out false alerts. For each registered impulsive signal, EEW systems need to decide (or, classify) in real-time whether or not the signal stems from an actual earthquake. State-of-the-art machine learning (ML) classifiers have been shown to strongly outperform more standard linear classifiers in a wide range of classification problems. Here we analyze the performance of a variety of different ML classifiers to identify which type of classifier leads to the most reliable signal/noise discrimination in an EEW context. We find that we can successfully train complex deep learning classifiers that can discriminate between nuisance and earthquake signals very reliably (accuracy of 99.5%). Less complex ML classifiers also outperform a linear classifier, but with significantly higher error rates. The deep ML classifiers may allow EEW systems to almost entirely avoid false and missed signal detections...

No abstract

No abstract

No abstract