The IJCAI-2021 awards were announced during the opening ceremony of the International Joint Conference on Artificial Intelligence (IJCAI-21). The honours included the 2021 AIJ classic paper award, the AIJ prominent paper award, and the IJCAI-JAIR best paper prize. This award recognizes outstanding papers, exceptional in their significance and impact, that were published at least 15 years ago, in the journal Artificial Intelligence (AIJ). This paper brought partially observable Markov decision processes (POMDPs) from the field of operational research to the field of AI. It provides an excellent account of the theory behind POMDPs, which demystified the field for a generation of researchers, and popularised their use in both AI and robotics.

In recent years a lot of research has been conducted within the area of causal inference and causal learning. Many methods have been developed to identify the cause-effect pairs in models and have been successfully applied to observational real-world data in order to determine the direction of causal relationships. Many of these methods require simplifying assumptions, such as absence of confounding, cycles, and selection bias. Yet in bivariate situations causal discovery problems remain challenging. One class of such methods, that also allows tackling the bivariate case, is based on Additive Noise Models (ANMs). Unfortunately, one aspect of these methods has not received much attention until now: what is the impact of different noise levels on the ability of these methods to identify the direction of the causal relationship. This work aims to bridge this gap with the help of an empirical study. For this work, we considered bivariate cases, which is the most elementary form of a causal discovery problem where one needs to decide whether X causes Y or Y causes X, given joint distributions of two variables X, Y. Furthermore, two specific methods have been selected, \textit{Regression with Subsequent Independence Test} and \textit{Identification using Conditional Variances}, which have been tested with an exhaustive range of ANMs where the additive noises' levels gradually change from 1% to 10000% of the causes' noise level (the latter remains fixed). Additionally, the experiments in this work consider several different types of distributions as well as linear and non-linear ANMs. The results of the experiments show that these methods can fail to capture the true causal direction for some levels of noise.

Increasing interest in privacy-preserving machine learning has led to new models for synthetic private data generation from undisclosed real data. However, mechanisms of privacy preservation introduce artifacts in the resulting synthetic data that have a significant impact on downstream tasks such as learning predictive models or inference. In particular, bias can affect all analyses as the synthetic data distribution is an inconsistent estimate of the real-data distribution. We propose several bias mitigation strategies using privatized likelihood ratios that have general applicability to differentially private synthetic data generative models. Through large-scale empirical evaluation, we show that bias mitigation provides simple and effective privacy-compliant augmentation for general applications of synthetic data. However, the work highlights that even after bias correction significant challenges remain on the usefulness of synthetic private data generators for tasks such as prediction and inference.

This paper identifies a structural property of data distributions that enables deep neural networks to learn hierarchically. We define the "staircase" property for functions over the Boolean hypercube, which posits that high-order Fourier coefficients are reachable from lower-order Fourier coefficients along increasing chains. We prove that functions satisfying this property can be learned in polynomial time using layerwise stochastic coordinate descent on regular neural networks -- a class of network architectures and initializations that have homogeneity properties. Our analysis shows that for such staircase functions and neural networks, the gradient-based algorithm learns high-level features by greedily combining lower-level features along the depth of the network. We further back our theoretical results with experiments showing that staircase functions are also learnable by more standard ResNet architectures with stochastic gradient descent. Both the theoretical and experimental results support the fact that staircase properties have a role to play in understanding the capabilities of gradient-based learning on regular networks, in contrast to general polynomial-size networks that can emulate any SQ or PAC algorithms as recently shown.

Central to the success of adaptive systems is their ability to interpret signals from their environment and respond accordingly -- they act as agents interacting with their surroundings. Such agents typically perform better when able to execute increasingly complex strategies. This comes with a cost: the more information the agent must recall from its past experiences, the more memory it will need. Here we investigate the power of agents capable of quantum information processing. We uncover the most general form a quantum agent need adopt to maximise memory compression advantages, and provide a systematic means of encoding their memory states. We show these encodings can exhibit extremely favourable scaling advantages relative to memory-minimal classical agents when information must be retained about events increasingly far into the past.

Describing the relationship between the variables in a study domain and modelling the data generating mechanism is a fundamental problem in many empirical sciences. Probabilistic graphical models are one common approach to tackle the problem. Learning the graphical structure is computationally challenging and a fervent area of current research with a plethora of algorithms being developed. To facilitate the benchmarking of different methods, we present a novel automated workflow, called benchpress for producing scalable, reproducible, and platform-independent benchmarks of structure learning algorithms for probabilistic graphical models. Benchpress is interfaced via a simple JSON-file, which makes it accessible for all users, while the code is designed in a fully modular fashion to enable researchers to contribute additional methodologies. Benchpress currently provides an interface to a large number of state-of-the-art algorithms from libraries such as BDgraph, BiDAG, bnlearn, GOBNILP, pcalg, r.blip, scikit-learn, TETRAD, and trilearn as well as a variety of methods for data generating models and performance evaluation. Alongside user-defined models and randomly generated datasets, the software tool also includes a number of standard datasets and graphical models from the literature, which may be included in a benchmarking workflow. We demonstrate the applicability of this workflow for learning Bayesian networks in four typical data scenarios. The source code and documentation is publicly available from http://github.com/felixleopoldo/benchpress.

Multimodal learning has achieved great successes in many scenarios. Compared with unimodal learning, it can effectively combine the information from different modalities to improve the performance of learning tasks. In reality, the multimodal data may have missing modalities due to various reasons, such as sensor failure and data transmission error. In previous works, the information of the modality-missing data has not been well exploited. To address this problem, we propose an efficient approach based on maximum likelihood estimation to incorporate the knowledge in the modality-missing data. Specifically, we design a likelihood function to characterize the conditional distribution of the modality-complete data and the modality-missing data, which is theoretically optimal. Moreover, we develop a generalized form of the softmax function to effectively implement maximum likelihood estimation in an end-to-end manner. Such training strategy guarantees the computability of our algorithm capably. Finally, we conduct a series of experiments on real-world multimodal datasets. Our results demonstrate the effectiveness of the proposed approach, even when 95% of the training data has missing modality.

To make advanced learning machines such as Deep Neural Networks (DNNs) more transparent in decision making, explainable AI (XAI) aims to provide interpretations of DNNs' predictions. These interpretations are usually given in the form of heatmaps, each one illustrating relevant patterns regarding the prediction for a given instance. Bayesian approaches such as Bayesian Neural Networks (BNNs) so far have a limited form of transparency (model transparency) already built-in through their prior weight distribution, but notably, they lack explanations of their predictions for given instances. In this work, we bring together these two perspectives of transparency into a holistic explanation framework for explaining BNNs. Within the Bayesian framework, the network weights follow a probability distribution. Hence, the standard (deterministic) prediction strategy of DNNs extends in BNNs to a predictive distribution, and thus the standard explanation extends to an explanation distribution. Exploiting this view, we uncover that BNNs implicitly employ multiple heterogeneous prediction strategies. While some of these are inherited from standard DNNs, others are revealed to us by considering the inherent uncertainty in BNNs. Our quantitative and qualitative experiments on toy/benchmark data and real-world data from pathology show that the proposed approach of explaining BNNs can lead to more effective and insightful explanations.

Automatic Speech Recognition (ASR) is an active field of research due to its huge number of applications and the proliferation of interfaces or computing devices that can support speech processing. But the bulk of applications is based on well-resourced languages that overshadow under-resourced ones. Yet ASR represents an undeniable mean to promote such languages, especially when design human-to-human or human-to-machine systems involving illiterate people. An approach to design an ASR system targeting under-resourced languages is to start with a limited vocabulary. ASR using a limited vocabulary is a subset of the speech recognition problem that focuses on the recognition of a small number of words or sentences. This paper aims to provide a comprehensive view of mechanisms behind ASR systems as well as techniques, tools, projects, recent contributions, and possibly future directions in ASR using a limited vocabulary. This work consequently provides a way to go when designing ASR system using limited vocabulary. Although an emphasis is put on limited vocabulary, most of the tools and techniques reported in this survey applied to ASR systems in general.

Cricket is unarguably one of the most popular sports in the world. Predicting the outcome of a cricket match has become a fundamental problem as we are advancing in the field of machine learning. Multiple researchers have tried to predict the outcome of a cricket match or a tournament, or to predict the performance of players during a match, or to predict the players who should be selected as per their current performance, form, morale, etc. using machine learning and artificial intelligence techniques keeping in mind extensive detailing, features, and parameters. We discuss some of these techniques along with a brief comparison among these techniques.