Autonomous cyber agents may be developed by applying reinforcement and deep reinforcement learning (RL/DRL), where agents are trained in a representative environment. The training environment must simulate with high-fidelity the network Cyber Operations (CyOp) that the agent aims to explore. Given the complexity of net-work CyOps, a good simulator is difficult to achieve. This work presents a systematic solution to automatically generate a high-fidelity simulator in the Cyber Gym for Intelligent Learning (CyGIL). Through representation learning and continuous learning, CyGIL provides a unified CyOp training environment where an emulated CyGIL-E automatically generates a simulated CyGIL-S. The simulator generation is integrated with the agent training process to further reduce the required agent training time. The agent trained in CyGIL-S is transferrable directly to CyGIL-E showing full transferability to the emulated "real" network. Experimental results are presented to demonstrate the CyGIL training performance. Enabling offline RL, the CyGIL solution presents a promising direction towards sim-to-real for leveraging RL agents in real-world cyber networks.

Pre-trained language models have attracted increasing attention in the biomedical domain, inspired by their great success in the general natural language domain. Among the two main branches of pre-trained language models in the general language domain, i.e., BERT (and its variants) and GPT (and its variants), the first one has been extensively studied in the biomedical domain, such as BioBERT and PubMedBERT. While they have achieved great success on a variety of discriminative downstream biomedical tasks, the lack of generation ability constrains their application scope. In this paper, we propose BioGPT, a domain-specific generative Transformer language model pre-trained on large scale biomedical literature. We evaluate BioGPT on six biomedical NLP tasks and demonstrate that our model outperforms previous models on most tasks. Especially, we get 44.98%, 38.42% and 40.76% F1 score on BC5CDR, KD-DTI and DDI end-to-end relation extraction tasks respectively, and 78.2% accuracy on PubMedQA, creating a new record. Our case study on text generation further demonstrates the advantage of BioGPT on biomedical literature to generate fluent descriptions for biomedical terms. Code is available at https://github.com/microsoft/BioGPT.

Task space trajectory tracking for quadruped robots plays a crucial role on achieving dexterous maneuvers in unstructured environments. To fulfill the control objective, the robot should apply forces through the contact of the legs with the supporting surface, while maintaining its stability and controllability. In order to ensure the operation of the robot under these conditions, one has to account for the possibility of unstable contact of the legs that arises when the robot operates on partially or globally slippery terrains. In this work, we propose an adaptive trajectory tracking controller for quadruped robots, which involves two prioritized layers of adaptation for avoiding possible slippage of one or multiple legs. The adaptive framework is evaluated through simulations and validated through experiments.

We modelled emergence and spread of the Omicron variant of SARS-CoV-2 in Australia between December 2021 and June 2022. This pandemic stage exhibited a diverse epidemiological profile with emergence of co-circulating sub-lineages of Omicron, further complicated by differences in social distancing behaviour which varied over time. Our study delineated distinct phases of the Omicron-associated pandemic stage, and retrospectively quantified the adoption of social distancing measures, fluctuating over different time periods in response to the observable incidence dynamics. We also modelled the corresponding disease burden, in terms of hospitalisations, intensive care unit occupancy, and mortality. Supported by good agreement between simulated and actual health data, our study revealed that the nonlinear dynamics observed in the daily incidence and disease burden were determined not only by introduction of sub-lineages of Omicron, but also by the fluctuating adoption of social distancing measures. Our high-resolution model can be used in design and evaluation of public health interventions during future crises.

In natural language processing, pre-trained language models have become essential infrastructures. However, these models often suffer from issues such as large size, long inference time, and challenging deployment. Moreover, most mainstream pre-trained models focus on English, and there are insufficient studies on small Chinese pre-trained models. In this paper, we introduce MiniRBT, a small Chinese pre-trained model that aims to advance research in Chinese natural language processing. MiniRBT employs a narrow and deep student model and incorporates whole word masking and two-stage distillation during pre-training to make it well-suited for most downstream tasks. Our experiments on machine reading comprehension and text classification tasks reveal that MiniRBT achieves 94% performance relative to RoBERTa, while providing a 6.8x speedup, demonstrating its effectiveness and efficiency.

In this paper, we propose an end-to-end Retrieval-Augmented Visual Language Model (REVEAL) that learns to encode world knowledge into a large-scale memory, and to retrieve from it to answer knowledge-intensive queries. REVEAL consists of four key components: the memory, the encoder, the retriever and the generator. The large-scale memory encodes various sources of multimodal world knowledge (e.g. image-text pairs, question answering pairs, knowledge graph triplets, etc) via a unified encoder. The retriever finds the most relevant knowledge entries in the memory, and the generator fuses the retrieved knowledge with the input query to produce the output. A key novelty in our approach is that the memory, encoder, retriever and generator are all pre-trained end-to-end on a massive amount of data. Furthermore, our approach can use a diverse set of multimodal knowledge sources, which is shown to result in significant gains. We show that REVEAL achieves state-of-the-art results on visual question answering and image captioning.

Well-established cognitive models coming from anthropology have shown that, due to the cognitive constraints that limit our "bandwidth" for social interactions, humans organize their social relations according to a regular structure. In this work, we postulate that similar regularities can be found in other cognitive processes, such as those involving language production. In order to investigate this claim, we analyse a dataset containing tweets of a heterogeneous group of Twitter users (regular users and professional writers). Leveraging a methodology similar to the one used to uncover the well-established social cognitive constraints, we find regularities at both the structural and semantic level. At the former, we find that a concentric layered structure (which we call ego network of words, in analogy to the ego network of social relationships) very well captures how individuals organise the words they use. The size of the layers in this structure regularly grows (approximately 2-3 times with respect to the previous one) when moving outwards, and the two penultimate external layers consistently account for approximately 60% and 30% of the used words, irrespective of the number of the total number of layers of the user. For the semantic analysis, each ring of each ego network is described by a semantic profile, which captures the topics associated with the words in the ring. We find that ring #1 has a special role in the model. It is semantically the most dissimilar and the most diverse among the rings. We also show that the topics that are important in the innermost ring also have the characteristic of being predominant in each of the other rings, as well as in the entire ego network. In this respect, ring #1 can be seen as the semantic fingerprint of the ego network of words.

The goal of data attribution is to trace model predictions back to training data. Despite a long line of work towards this goal, existing approaches to data attribution tend to force users to choose between computational tractability and efficacy. That is, computationally tractable methods can struggle with accurately attributing model predictions in non-convex settings (e.g., in the context of deep neural networks), while methods that are effective in such regimes require training thousands of models, which makes them impractical for large models or datasets. In this work, we introduce TRAK (Tracing with the Randomly-projected After Kernel), a data attribution method that is both effective and computationally tractable for large-scale, differentiable models. In particular, by leveraging only a handful of trained models, TRAK can match the performance of attribution methods that require training thousands of models. We demonstrate the utility of TRAK across various modalities and scales: image classifiers trained on ImageNet, vision-language models (CLIP), and language models (BERT and mT5). We provide code for using TRAK (and reproducing our work) at https://github.com/MadryLab/trak .

As a second-order method, the Natural Gradient Descent (NGD) has the ability to accelerate training of neural networks. However, due to the prohibitive computational and memory costs of computing and inverting the Fisher Information Matrix (FIM), efficient approximations are necessary to make NGD scalable to Deep Neural Networks (DNNs). Many such approximations have been attempted. The most sophisticated of these is KFAC, which approximates the FIM as a block-diagonal matrix, where each block corresponds to a layer of the neural network. By doing so, KFAC ignores the interactions between different layers. In this work, we investigate the interest of restoring some low-frequency interactions between the layers by means of two-level methods. Inspired from domain decomposition, several two-level corrections to KFAC using different coarse spaces are proposed and assessed. The obtained results show that incorporating the layer interactions in this fashion does not really improve the performance of KFAC. This suggests that it is safe to discard the off-diagonal blocks of the FIM, since the block-diagonal approach is sufficiently robust, accurate and economical in computation time.

Sea surface temperature (SST) forecasts help with managing the marine ecosystem and the aquaculture impacted by anthropogenic climate change. Numerical dynamical models are resource intensive for SST forecasts; machine learning (ML) models could reduce high computational requirements and have been in the focus of the research community recently. ML models normally require a large amount of data for training. Environmental data are collected on regularly-spaced grids, so early work mainly used grid-based deep learning (DL) for prediction. However, both grid data and the corresponding DL approaches have inherent problems. As geometric DL has emerged, graphs as a more generalized data structure and graph neural networks (GNNs) have been introduced to the spatiotemporal domains. In this work, we preliminarily explored graph re-sampling and GNNs for global SST forecasts, and GNNs show better one month ahead SST prediction than the persistence model in most oceans in terms of root mean square errors.