Ensembles of geophysical models improve prediction accuracy and express uncertainties. We develop a novel data-driven ensembling strategy for combining geophysical models using Bayesian Neural Networks, which infers spatiotemporally varying model weights and bias, while accounting for heteroscedastic uncertainties in the observations. This produces more accurate and uncertaintyaware predictions without sacrificing interpretability. Applied to the prediction of total column ozone from an ensemble of 15 chemistry-climate models, we find that the Bayesian neural network ensemble (BayNNE) outperforms existing methods for ensembling physical models, achieving a 49.4% reduction in RMSE for temporal extrapolation, and a 67.4% reduction in RMSE for polar data voids, compared to a weighted mean. Uncertainty is also well-characterized, with 91.9% of the data points in our extrapolation validation dataset lying within 2 standard deviations and 98.9% within 3 standard deviations.

Existing approaches to federated learning suffer from a communication bottleneck as well as convergence issues due to sparse client participation. In this paper we introduce a novel algorithm, called FetchSGD, to overcome these challenges. FetchSGD compresses model updates using a Count Sketch, and then takes advantage of the mergeability of sketches to combine model updates from many workers. A key insight in the design of FetchSGD is that, because the Count Sketch is linear, momentum and error accumulation can both be carried out within the sketch. This allows the algorithm to move momentum and error accumulation from clients to the central aggregator, overcoming the challenges of sparse client participation while still achieving high compression rates and good convergence. We prove that FetchSGD has favorable convergence guarantees, and we demonstrate its empirical effectiveness by training two residual networks and a transformer model.

Many successful deep learning architectures are equivariant to certain transformations in order to conserve parameters and improve generalization: most famously, convolution layers are equivariant to shifts of the input. This approach only works when practitioners know the symmetries of the task and can manually construct an architecture with the corresponding equivariances. Our goal is an approach for learning equivariances from data, without needing to design custom task-specific architectures. We present a method for learning and encoding equivariances into networks by learning corresponding parameter sharing patterns from data. Our method can provably represent equivariance-inducing parameter sharing for any finite group of symmetry transformations. Our experiments suggest that it can automatically learn to encode equivariances to common transformations used in image processing tasks. We provide our experiment code at https://github.com/AllanYangZhou/metalearning-symmetries.

Noname manuscript No. (will be inserted by the editor) Abstract Bayesian Additive Regression Trees (BART) 1 Introduction is a tree-based machine learning method that has been successfully applied to regression and classification problems. Bayesian Additive Regression Trees (BART) is a statistical BART assumes regularisation priors on a set of method proposed by Chipman et al (2010) that has trees that work as weak learners and is very flexible for become popular in recent years due to its competitive predicting in the presence of non-linearity and highorder performance on regression and classification problems, interactions. In this paper, we introduce an extension when compared to other supervised machine learning of BART, called Model Trees BART (MOTR-methods, such as Random Forests (RF) (Breiman, 2001) BART), that considers piecewise linear functions at node and Gradient Boosting (GB) (Friedman, 2001). In MOTR-BART, differs from other tree-based methods as it controls the rather than having a unique value at node level for the structure of each tree via a prior distribution and generates prediction, a linear predictor is estimated considering the predictions via an MCMC backfitting algorithm the covariates that have been used as the split variables that is responsible for accepting and rejecting the in the corresponding tree. In our approach, local linearities proposed trees along the iterations.

Transformer models have garnered a lot of interest in recent years by delivering state-of-the-art performance in a range of Natural Language Processing (NLP) tasks. However, these models can have over a hundred billion parameters, presenting very high computational and memory requirements. We address this challenge through Approximate Computing, specifically targeting the use of Transformers in NLP tasks. Transformers are typically pre-trained and subsequently specialized for specific tasks through transfer learning. Based on the observation that pre-trained Transformers are often over-parameterized for several downstream NLP tasks, we propose a framework to create smaller, faster and in some cases more accurate models. The key cornerstones of the framework are a Significance Analysis (SA) method that identifies components in a pre-trained Transformer that are less significant for a given task, and techniques to approximate the less significant components. Our approximations include pruning of blocks, attention heads and weight groups, quantization of less significant weights and a low-complexity sign-matching based attention mechanism. Our framework can be adapted to produce models that are faster, smaller and/or more accurate, depending on the user's constraints. We apply our framework to seven Transformer models, including optimized models like DistilBERT and Q8BERT, and three downstream tasks. We demonstrate that our framework produces models that are up to 4 faster and up to 14 smaller (with less than 0.5% relative accuracy degradation), or up to 5.5% more accurate with simultaneous improvements of up to 9.83 in model size or 2.94 in speed. Transformer networks with hundreds of billions of parameters, such as T5 ([17]), Megatron ([22]), BERT ([2]), GPT-2 ( [16]) and GPT-3 ([1]), have achieved state-of-the-art performance in several Natural Language Processing tasks. Model sizes are expected to grow further in the future as increasing the number of parameters has been shown to improve performance.

We present a method for learning representations of entities, that uses a Transformer-based architecture as an entity encoder, and link prediction training on a knowledge graph with textual entity descriptions. We demonstrate that our approach can be applied effectively for link prediction in different inductive settings involving entities not seen during training, outperforming related state-of-the-art methods (22% MRR improvement on average). We provide evidence that the learned representations transfer to other tasks that do not require fine-tuning the entity encoder. In an entity classification task we obtain an average improvement of 16% accuracy compared with baselines that also employ pre-trained models. For an information retrieval task, significant improvements of up to 8.8% in NDCG@10 were obtained for natural language queries.

Over the next few years, society as a whole will need to address what core values it wishes to protect when dealing with technology. Anthropology, a field dedicated to the very notion of what it means to be human, can provide some interesting insights into how to cope and tackle these changes in our Western society and other areas of the world. It can be challenging for social science practitioners to grasp and keep up with the pace of technological innovation, with many being unfamiliar with the jargon of AI. This short guide serves as both an introduction to AI ethics and social science and anthropological perspectives on the development of AI. It intends to provide those unfamiliar with the field with an insight into the societal impact of AI systems and how, in turn, these systems can lead us to rethink how our world operates.

Past work on story generation has demonstrated the usefulness of conditioning on a generation plan to generate coherent stories. However, these approaches have used heuristics or off-the-shelf models to first tag training stories with the desired type of plan, and then train generation models in a supervised fashion. In this paper, we propose a deep latent variable model that first samples a sequence of anchor words, one per sentence in the story, as part of its generative process. During training, our model treats the sequence of anchor words as a latent variable and attempts to induce anchoring sequences that help guide generation in an unsupervised fashion. We conduct experiments with several types of sentence decoder distributions: left-to-right and non-monotonic, with different degrees of restriction. Further, since we use amortized variational inference to train our model, we introduce two corresponding types of inference network for predicting the posterior on anchor words. We conduct human evaluations which demonstrate that the stories produced by our model are rated better in comparison with baselines which do not consider story plans, and are similar or better in quality relative to baselines which use external supervision for plans. Additionally, the proposed model gets favorable scores when evaluated on perplexity, diversity, and control of story via discrete plan.

Daniel Faggella is Head of Research at Emerj. Called upon by the United Nations, World Bank, INTERPOL, and leading enterprises, Daniel is a globally sought-after expert on the competitive strategy implications of AI for business and government leaders. Companies and cities all over world are experimenting with using artificial intelligence to reduce and prevent crime, and to more quickly respond to crimes in progress. The ideas behind many of these projects is that crimes are relatively predictable; it just requires being able to sort through a massive volume of data to find patterns that are useful to law enforcement. This kind of data analysis was technologically impossible a few decades ago, but the hope is that recent developments in machine learning are up to the task.

We propose a method to automatically generate a domain- and task-adaptive maskings of the given text for self-supervised pre-training, such that we can effectively adapt the language model to a particular target task (e.g. question answering). Specifically, we present a novel reinforcement learning-based framework which learns the masking policy, such that using the generated masks for further pre-training of the target language model helps improve task performance on unseen texts. We use off-policy actor-critic with entropy regularization and experience replay for reinforcement learning, and propose a Transformer-based policy network that can consider the relative importance of words in a given text. We validate our Neural Mask Generator (NMG) on several question answering and text classification datasets using BERT and DistilBERT as the language models, on which it outperforms rule-based masking strategies, by automatically learning optimal adaptive maskings.