Imitation Learning (IL) is one of the most widely used methods in machine learning. Yet, while powerful, many works find it is often not able to fully recover the underlying expert behavior [1-3]. However, none of these works deeply investigate the role of scaling up the model and data size. Inspired by recent work in Natural Language Processing (NLP) [4, 5] where "scaling up" has resulted in increasingly more capable LLMs, we investigate whether carefully scaling up model and data size can bring similar improvements in the imitation learning setting. To demonstrate our findings, we focus on the game of NetHack, a challenging environment featuring procedural generation, stochasticity, long-term dependencies, and partial observability. We find IL loss and mean return scale smoothly with the compute budget and are strongly correlated, resulting in power laws for training compute-optimal IL agents with respect to model size and number of samples. We forecast and train several NetHack agents with IL and find they outperform prior state-of-the-art by at least 2x in all settings. Our work both demonstrates the scaling behavior of imitation learning in a challenging domain, as well as the viability of scaling up current approaches for increasingly capable agents in NetHack, a game that remains elusively hard for current AI systems.

In many reinforcement learning tasks, the agent has to learn to interact with many objects of different types and generalize to unseen combinations and numbers of objects. Often a task is a composition of previously learned tasks (e.g. block stacking). These are examples of compositional generalization, in which we compose object-centric representations to solve complex tasks. Recent works have shown the benefits of object-factored representations and hierarchical abstractions for improving sample efficiency in these settings. On the other hand, these methods do not fully exploit the benefits of factorization in terms of object attributes. In this paper, we address this opportunity and introduce the Dynamic Attribute FacTored RL (DAFT-RL) framework. In DAFT-RL, we leverage object-centric representation learning to extract objects from visual inputs. We learn to classify them in classes and infer their latent parameters. For each class of object, we learn a class template graph that describes how the dynamics and reward of an object of this class factorize according to its attributes. We also learn an interaction pattern graph that describes how objects of different classes interact with each other at the attribute level. Through these graphs and a dynamic interaction graph that models the interactions between objects, we can learn a policy that can then be directly applied in a new environment by just estimating the interactions and latent parameters. We evaluate DAFT-RL in three benchmark datasets and show our framework outperforms the state-of-the-art in generalizing across unseen objects with varying attributes and latent parameters, as well as in the composition of previously learned tasks.

We propose a general framework for decoding quantum error-correcting codes with generative modeling. The model utilizes autoregressive neural networks, specifically Transformers, to learn the joint probability of logical operators and syndromes. This training is in an unsupervised way, without the need for labeled training data, and is thus referred to as pre-training. After the pre-training, the model can efficiently compute the likelihood of logical operators for any given syndrome, using maximum likelihood decoding. It can directly generate the most-likely logical operators with computational complexity $\mathcal O(2k)$ in the number of logical qubits $k$, which is significantly better than the conventional maximum likelihood decoding algorithms that require $\mathcal O(4^k)$ computation. Based on the pre-trained model, we further propose refinement to achieve more accurately the likelihood of logical operators for a given syndrome by directly sampling the stabilizer operators. We perform numerical experiments on stabilizer codes with small code distances, using both depolarizing error models and error models with correlated noise. The results show that our approach provides significantly better decoding accuracy than the minimum weight perfect matching and belief-propagation-based algorithms. Our framework is general and can be applied to any error model and quantum codes with different topologies such as surface codes and quantum LDPC codes. Furthermore, it leverages the parallelization capabilities of GPUs, enabling simultaneous decoding of a large number of syndromes. Our approach sheds light on the efficient and accurate decoding of quantum error-correcting codes using generative artificial intelligence and modern computational power.

Text generative models trained via Maximum Likelihood Estimation (MLE) suffer from the notorious exposure bias problem, and Generative Adversarial Networks (GANs) are shown to have potential to tackle this problem. Existing language GANs adopt estimators like REINFORCE or continuous relaxations to model word probabilities. The inherent limitations of such estimators lead current models to rely on pre-training techniques (MLE pre-training or pre-trained embeddings). Representation modeling methods which are free from those limitations, however, are seldomly explored because of their poor performance in previous attempts. Our analyses reveal that invalid sampling methods and unhealthy gradients are the main contributors to such unsatisfactory performance. In this work, we present two techniques to tackle these problems: dropout sampling and fully normalized LSTM. Based on these two techniques, we propose InitialGAN whose parameters are randomly initialized in full. Besides, we introduce a new evaluation metric, Least Coverage Rate, to better evaluate the quality of generated samples. The experimental results demonstrate that InitialGAN outperforms both MLE and other compared models. To the best of our knowledge, it is the first time a language GAN can outperform MLE without using any pre-training techniques.

In recent years, word embeddings have been widely used to measure biases in texts. Even if they have proven to be effective in detecting a wide variety of biases, metrics based on word embeddings lack transparency and interpretability. We analyze an alternative PMI-based metric to quantify biases in texts. It can be expressed as a function of conditional probabilities, which provides a simple interpretation in terms of word co-occurrences. We also prove that it can be approximated by an odds ratio, which allows estimating confidence intervals and statistical significance of textual biases. This approach produces similar results to metrics based on word embeddings when capturing gender gaps of the real world embedded in large corpora.

Cyber threat attribution is the process of identifying the actor of an attack incident in cyberspace. An accurate and timely threat attribution plays an important role in deterring future attacks by applying appropriate and timely defense mechanisms. Manual analysis of attack patterns gathered by honeypot deployments, intrusion detection systems, firewalls, and via trace-back procedures is still the preferred method of security analysts for cyber threat attribution. Such attack patterns are low-level Indicators of Compromise (IOC). They represent Tactics, Techniques, Procedures (TTP), and software tools used by the adversaries in their campaigns. The adversaries rarely re-use them. They can also be manipulated, resulting in false and unfair attribution. To empirically evaluate and compare the effectiveness of both kinds of IOC, there are two problems that need to be addressed. The first problem is that in recent research works, the ineffectiveness of low-level IOC for cyber threat attribution has been discussed intuitively. An empirical evaluation for the measure of the effectiveness of low-level IOC based on a real-world dataset is missing. The second problem is that the available dataset for high-level IOC has a single instance for each predictive class label that cannot be used directly for training machine learning models. To address these problems in this research work, we empirically evaluate the effectiveness of low-level IOC based on a real-world dataset that is specifically built for comparative analysis with high-level IOC. The experimental results show that the high-level IOC trained models effectively attribute cyberattacks with an accuracy of 95% as compared to the low-level IOC trained models where accuracy is 40%.

Effective and efficient malware detection is at the forefront of research into building secure digital systems. As with many other fields, malware detection research has seen a dramatic increase in the application of machine learning algorithms. One machine learning technique that has been used widely in the field of pattern matching in general-and malware detection in particular-is hidden Markov models (HMMs). HMM training is based on a hill climb, and hence we can often improve a model by training multiple times with different initial values. In this research, we compare boosted HMMs (using AdaBoost) to HMMs trained with multiple random restarts, in the context of malware detection. These techniques are applied to a variety of challenging malware datasets. We find that random restarts perform surprisingly well in comparison to boosting. Only in the most difficult "cold start" cases (where training data is severely limited) does boosting appear to offer sufficient improvement to justify its higher computational cost in the scoring phase.

Algorithmic and data-driven decisions and recommendations are commonly used in high-stakes decision-making settings such as criminal justice, medicine, and public policy. We investigate whether it would have been possible to improve a security assessment algorithm employed during the Vietnam War, using outcomes measured immediately after its introduction in late 1969. This empirical application raises several methodological challenges that frequently arise in high-stakes algorithmic decision-making. First, before implementing a new algorithm, it is essential to characterize and control the risk of yielding worse outcomes than the existing algorithm. Second, the existing algorithm is deterministic, and learning a new algorithm requires transparent extrapolation. Third, the existing algorithm involves discrete decision tables that are common but difficult to optimize over. To address these challenges, we introduce the Average Conditional Risk (ACRisk), which first quantifies the risk that a new algorithmic policy leads to worse outcomes for subgroups of individual units and then averages this over the distribution of subgroups. We also propose a Bayesian policy learning framework that maximizes the posterior expected value while controlling the posterior expected ACRisk. This framework separates the estimation of heterogeneous treatment effects from policy optimization, enabling flexible estimation of effects and optimization over complex policy classes. We characterize the resulting chance-constrained optimization problem as a constrained linear programming problem. Our analysis shows that compared to the actual algorithm used during the Vietnam War, the learned algorithm assesses most regions as more secure and emphasizes economic and political factors over military factors.

In multi-timescale multi-agent reinforcement learning (MARL), agents interact across different timescales. In general, policies for time-dependent behaviors, such as those induced by multiple timescales, are non-stationary. Learning non-stationary policies is challenging and typically requires sophisticated or inefficient algorithms. Motivated by the prevalence of this control problem in real-world complex systems, we introduce a simple framework for learning non-stationary policies for multi-timescale MARL. Our approach uses available information about agent timescales to define a periodic time encoding. In detail, we theoretically demonstrate that the effects of non-stationarity introduced by multiple timescales can be learned by a periodic multi-agent policy. To learn such policies, we propose a policy gradient algorithm that parameterizes the actor and critic with phase-functioned neural networks, which provide an inductive bias for periodicity. The framework's ability to effectively learn multi-timescale policies is validated on a gridworld and building energy management environment.

In this white paper, we synthesize key points made during presentations and discussions from the AI-Assisted Decision Making for Conservation workshop, hosted by the Center for Research on Computation and Society at Harvard University on October 20-21, 2022. We identify key open research questions in resource allocation, planning, and interventions for biodiversity conservation, highlighting conservation challenges that not only require AI solutions, but also require novel methodological advances. In addition to providing a summary of the workshop talks and discussions, we hope this document serves as a call-to-action to orient the expansion of algorithmic decision-making approaches to prioritize real-world conservation challenges, through collaborative efforts of ecologists, conservation decision-makers, and AI researchers.