If you follow my blogs, you know that I've been focusing a fair amount of attention on artificial intelligence, and how it has raised reasons for both optimism and extreme ethical pause. In this one, I want to discuss how there is potential for a new conflict not dissimilar to the Cold War with the development and proliferation of nuclear energy; but this time AI will take centre stage of the theatre. Very much akin to nuclear expansion, artificial intelligence comes with its own bag of pros and cons. Indubitably, nuclear energy has been harnessed for the commonwealth of mankind. Water Desalination -- Reducing the saline content of seawater is extremely costly and inefficient.

Traditional cybersecurity isn't necessarily bad at detecting attacks, the trouble is it often does so after they have occurred. A better approach is to spot potential attacks and block them before they can do any damage. One possible way of doing this is via'deep learning' allowing technology to identify the difference between good and bad. We spoke with Brooks Wallace, cybersecurity sales leader at Deep Instinct to find out more about this innovative solution. BW: If you look at cybersecurity, there's always been this holy grail of prevention.

Extreme polarization can undermine democracy by making compromise impossible and transforming politics into a zero-sum game. Ideological polarization - the extent to which political views are widely dispersed - is already strong among elites, but less so among the general public (McCarty, 2019, p. 50-68). Strong mutual distrust and hostility between Democrats and Republicans in the U.S., combined with the elites' already strong ideological polarization, could lead to increasing ideological polarization among the public. The paper addresses two questions: (1) Is there a level of ideological polarization above which polarization feeds upon itself to become a runaway process? (2) If so, what policy interventions could prevent such dangerous positive feedback loops? To explore these questions, we present an agent-based model of ideological polarization that differentiates between the tendency for two actors to interact (exposure) and how they respond when interactions occur, positing that interaction between similar actors reduces their difference while interaction between dissimilar actors increases their difference. Our analysis explores the effects on polarization of different levels of tolerance to other views, responsiveness to other views, exposure to dissimilar actors, multiple ideological dimensions, economic self-interest, and external shocks. The results suggest strategies for preventing, or at least slowing, the development of extreme polarization.

The digital media, identified as computational propaganda provides a pathway for propaganda to expand its reach without limit. State-backed propaganda aims to shape the audiences' cognition toward entities in favor of a certain political party or authority. Furthermore, it has become part of modern information warfare used in order to gain an advantage over opponents. Most of the current studies focus on using machine learning, quantitative, and qualitative methods to distinguish if a certain piece of information on social media is propaganda. Mainly conducted on English content, but very little research addresses Chinese Mandarin content. From propaganda detection, we want to go one step further to provide more fine-grained information on propaganda techniques that are applied. In this research, we aim to bridge the information gap by providing a multi-labeled propaganda techniques dataset in Mandarin based on a state-backed information operation dataset provided by Twitter. In addition to presenting the dataset, we apply a multi-label text classification using fine-tuned BERT. Potentially this could help future research in detecting state-backed propaganda online especially in a cross-lingual context and cross platforms identity consolidation.

Node classification and link prediction are widely studied tasks in graph representation learning. While both transductive node classification and link prediction operate over a single input graph, they are studied in isolation so far, which leads to discrepancies. Node classification models take as input a graph with node features and incomplete node labels, and implicitly assume that the input graph is relationally complete, i.e., no edges are missing from the input graph. This is in sharp contrast with link prediction models that are solely motivated by the relational incompleteness of the input graph which does not have any node features. We propose a unifying perspective and study the problems of (i) transductive node classification over incomplete graphs and (ii) link prediction over graphs with node features. We propose an extension to an existing box embedding model, and show that this model is fully expressive, and can solve both of these tasks in an end-to-end fashion. To empirically evaluate our model, we construct a knowledge graph with node features, which is challenging both for node classification and link prediction. Our model performs very strongly when compared to the respective state-of-the-art models for node classification and link prediction on this dataset and shows the importance of a unified perspective for node classification and link prediction on knowledge graphs.

Explainable Artificial Intelligence (XAI) is a set of techniques that allows the understanding of both technical and non-technical aspects of Artificial Intelligence (AI) systems. XAI is crucial to help satisfying the increasingly important demand of \emph{trustworthy} Artificial Intelligence, characterized by fundamental characteristics such as respect of human autonomy, prevention of harm, transparency, accountability, etc. Within XAI techniques, counterfactual explanations aim to provide to end users a set of features (and their corresponding values) that need to be changed in order to achieve a desired outcome. Current approaches rarely take into account the feasibility of actions needed to achieve the proposed explanations, and in particular they fall short of considering the causal impact of such actions. In this paper, we present Counterfactual Explanations as Interventions in Latent Space (CEILS), a methodology to generate counterfactual explanations capturing by design the underlying causal relations from the data, and at the same time to provide feasible recommendations to reach the proposed profile. Moreover, our methodology has the advantage that it can be set on top of existing counterfactuals generator algorithms, thus minimising the complexity of imposing additional causal constrains. We demonstrate the effectiveness of our approach with a set of different experiments using synthetic and real datasets (including a proprietary dataset of the financial domain).

As machine learning (ML) systems take a more prominent and central role in contributing to life-impacting decisions, ensuring their trustworthiness and accountability is of utmost importance. Explanations sit at the core of these desirable attributes of a ML system. The emerging field is frequently called ``Explainable AI (XAI)'' or ``Explainable ML.'' The goal of explainable ML is to intuitively explain the predictions of a ML system, while adhering to the needs to various stakeholders. Many explanation techniques were developed with contributions from both academia and industry. However, there are several existing challenges that have not garnered enough interest and serve as roadblocks to widespread adoption of explainable ML. In this short paper, we enumerate challenges in explainable ML from an industry perspective. We hope these challenges will serve as promising future research directions, and would contribute to democratizing explainable ML.

Restless and collapsing bandits are commonly used to model constrained resource allocation in settings featuring arms with action-dependent transition probabilities, such as allocating health interventions among patients [Whittle, 1988; Mate et al., 2020]. However, state-of-the-art Whittle-index-based approaches to this planning problem either do not consider fairness among arms, or incentivize fairness without guaranteeing it [Mate et al., 2021]. Additionally, their optimality guarantees only apply when arms are indexable and threshold-optimal. We demonstrate that the incorporation of hard fairness constraints necessitates the coupling of arms, which undermines the tractability, and by extension, indexability of the problem. We then introduce ProbFair, a probabilistically fair stationary policy that maximizes total expected reward and satisfies the budget constraint, while ensuring a strictly positive lower bound on the probability of being pulled at each timestep. We evaluate our algorithm on a real-world application, where interventions support continuous positive airway pressure (CPAP) therapy adherence among obstructive sleep apnea (OSA) patients, as well as simulations on a broader class of synthetic transition matrices.

We study transfer learning in the presence of spurious correlations. We experimentally demonstrate that directly transferring the stable feature extractor learned on the source task may not eliminate these biases for the target task. However, we hypothesize that the unstable features in the source task and those in the target task are directly related. By explicitly informing the target classifier of the source task's unstable features, we can regularize the biases in the target task. Specifically, we derive a representation that encodes the unstable features by contrasting different data environments in the source task. On the target task, we cluster data from this representation, and achieve robustness by minimizing the worst-case risk across all clusters. We evaluate our method on both text and image classifications. Empirical results demonstrate that our algorithm is able to maintain robustness on the target task, outperforming the best baseline by 22.9% in absolute accuracy across 12 transfer settings. Our code is available at https://github.com/YujiaBao/Tofu.

Recent studies have exposed that many graph neural networks (GNNs) are sensitive to adversarial attacks, and can suffer from performance loss if the graph structure is intentionally perturbed. A different line of research has shown that many GNN architectures implicitly assume that the underlying graph displays homophily, i.e., connected nodes are more likely to have similar features and class labels, and perform poorly if this assumption is not fulfilled. In this work, we formalize the relation between these two seemingly different issues. We theoretically show that in the standard scenario in which node features exhibit homophily, impactful structural attacks always lead to increased levels of heterophily. Then, inspired by GNN architectures that target heterophily, we present two designs -- (i) separate aggregators for ego- and neighbor-embeddings, and (ii) a reduced scope of aggregation -- that can significantly improve the robustness of GNNs. Our extensive empirical evaluations show that GNNs featuring merely these two designs can achieve significantly improved robustness compared to the best-performing unvaccinated model with 24.99% gain in average performance under targeted attacks, while having smaller computational overhead than existing defense mechanisms. Furthermore, these designs can be readily combined with explicit defense mechanisms to yield state-of-the-art robustness with up to 18.33% increase in performance under attacks compared to the best-performing vaccinated model.