AAAI 2023 Papers

Despite that convolution neural networks (CNN) have recently demonstrated high-quality reconstruction for video super-resolution (VSR), efficiently training competitive VSR models remains a challenging problem. It usually takes an order of magnitude more time than training their counterpart image models, leading to long research cycles. Existing VSR methods typically train models with fixed spatial and temporal sizes from beginning to end. The fixed sizes are usually set to large values for good performance, resulting to slow training. However, is such a rigid training strategy necessary for VSR? In this work, we show that it is possible to gradually train video models from small to large spatial/temporal sizes, \ie, in an easy-to-hard manner. In particular, the whole training is divided into several stages and the earlier stage has smaller training spatial shape. Inside each stage, the temporal size also varies from short to long while the spatial size remains unchanged. Training is accelerated by such a multigrid training strategy, as most of computation is performed on smaller spatial and shorter temporal shapes. For further acceleration with GPU parallelization, we also investigate the large minibatch training without the loss in accuracy. Extensive experiments demonstrate that our method is capable of largely speeding up training (up to $6.2\times$ speedup in wall-clock training time) without performance drop for various VSR models.

In the scope of "AI for Science", solving inverse problems is a longstanding challenge in materials and drug discovery, where the goal is to determine the hidden structures given a set of desirable properties. Deep generative models are recently proposed to solve inverse problems, but these are currently struggling in expensive forward operators, precisely localizing the exact solutions and fully exploring the parameter spaces without missing solutions. In this work, we propose a novel approach (called iPage) to accelerate the inverse learning process by leveraging probabilistic inference from deep invertible models and deterministic optimization via fast gradient descent. Given a target property, the learned invertible model provides a posterior over the parameter space; we identify these posterior samples as an intelligent prior initialization which enables us to narrow down the search space. We then perform gradient descent to calibrate the inverse solutions within a local region. Meanwhile, a space-filling sampling is imposed on the latent space to better explore and capture all possible solutions. We evaluate our approach on three benchmark tasks and create two datasets of real-world applications from quantum chemistry and additive manufacturing and find our method achieves superior performance compared to several state-of-the-art baseline methods. The iPage code is available at https://github.com/jxzhangjhu/MatDesINNe.

Dung's abstract Argumentation Framework (AF) has emerged as a central formalism in the area of knowledge representation and reasoning. Preferences in AF allow to represent the comparative strength of arguments in a simple yet expressive way. Preference-based AF (PAF) has been proposed to extend AF with preferences of the form a > b, whose intuitive meaning is that argument a is better than b. In this paper we generalize PAF by introducing conditional preferences of the form a > b \leftarrow body that informally state that a is better than b whenever the condition expressed by body is true. The resulting framework, namely Conditional Preference-based AF (CPAF), extends the PAF semantics under three well-known preference criteria, i.e. democratic, elitist, and KTV. After introducing CPAF, we study the complexity of the verification problem (deciding whether a set of arguments is a ``best'' extension) as well as of the credulous and skeptical acceptance problems (deciding whether a given argument belongs to any or all ``best'' extensions, respectively) under multiple-status semantics (that is, complete, preferred, stable, and semi-stable semantics) for the above-mentioned preference criteria.

We introduce AVCAffe, the first Audio-Visual dataset consisting of Cognitive load and Affect attributes. We record AVCAffe by simulating remote work scenarios over a video-conferencing platform, where subjects collaborate to complete a number of cognitively engaging tasks. AVCAffe is the largest originally collected (not collected from the Internet) affective dataset in English language. We recruit 106 participants from 18 different countries of origin, spanning an age range of 18 to 57 years old, with a balanced male-female ratio. AVCAffe comprises a total of 108 hours of video, equivalent to more than 58,000 clips along with task-based self-reported ground truth labels for arousal, valence, and cognitive load attributes such as mental demand, temporal demand, effort, and a few others. We believe AVCAffe would be a challenging benchmark for the deep learning research community given the inherent difficulty of classifying affect and cognitive load in particular. Moreover, our dataset fills an existing timely gap by facilitating the creation of learning systems for better self-management of remote work meetings, and further study of hypotheses regarding the impact of remote work on cognitive load and affective states.

Current work in named entity recognition (NER) uses either cross entropy (CE) or conditional random fields (CRF) as the objective/loss functions to optimize the underlying NER model. Both of these traditional objective functions for the NER problem generally produce adequate performance when the data distribution is balanced and there are sufficient annotated training examples. But since NER is inherently an imbalanced tagging problem, the model performance under the low-resource settings could suffer using these standard objective functions. Based on recent advances in area under the ROC curve (AUC) maximization, we propose to optimize the NER model by maximizing the AUC score. We give evidence that by simply combining two binary-classifiers that maximize the AUC score, significant performance improvement over traditional loss functions is achieved under low-resource NER settings. We also conduct extensive experiments to demonstrate the advantages of our method under the low-resource and highly-imbalanced data distribution settings. To the best of our knowledge, this is the first work that brings AUC maximization to the NER setting. Furthermore, we show that our method is agnostic to different types of NER embeddings, models and domains. The code of this work is available at https://github.com/dngu0061/NER-AUC-2T.

Transformer-based autoregressive (AR) methods have achieved appealing performance for varied sequence-to-sequence generation tasks, e.g., neural machine translation, summarization, and code generation, but suffer from low inference efficiency. To speed up the inference stage, many non-autoregressive (NAR) strategies have been proposed in the past few years. Among them, the conditional masked language model (CMLM) is one of the most versatile frameworks, as it can support many different sequence generation scenarios and achieve very competitive performance on these tasks. In this paper, we further introduce a simple yet effective adaptive masking over masking strategy to enhance the refinement capability of the decoder and make the encoder optimization easier. Experiments on 3 different tasks (neural machine translation, summarization, and code generation) with 15 datasets in total confirm that our proposed simple method achieves significant performance improvement over the strong CMLM model. Surprisingly, our proposed model yields state-of-the-art performance on neural machine translation (34.62 BLEU on WMT16 EN to RO, 34.82 BLEU on WMT16 RO to EN, and 34.84 BLEU on IWSLT De to En) and even better performance than the AR Transformer on 7 benchmark datasets with at least 2.2x speedup. Our code is available at GitHub.

Evaluating neural network performance is critical to deep neural network design but a costly procedure. Neural predictors provide an efficient solution by treating architectures as samples and learning to estimate their performance on a given task. However, existing predictors are task-dependent, predominantly estimating neural network performance on image classification benchmarks. They are also search-space dependent; each predictor is designed to make predictions for a specific architecture search space with predefined topologies and set of operations. In this paper, we propose a novel All-in-One Predictor (AIO-P), which aims to pretrain neural predictors on architecture examples from multiple, separate computer vision (CV) task domains and multiple architecture spaces, and then transfer to unseen downstream CV tasks or neural architectures. We describe our proposed techniques for general graph representation, efficient predictor pretraining and knowledge infusion techniques, as well as methods to transfer to downstream tasks/spaces. Extensive experimental results show that AIO-P can achieve Mean Absolute Error (MAE) and Spearman’s Rank Correlation (SRCC) below 1p% and above 0.5, respectively, on a breadth of target downstream CV tasks with or without fine-tuning, outperforming a number of baselines. Moreover, AIO-P can directly transfer to new architectures not seen during training, accurately rank them and serve as an effective performance estimator when paired with an algorithm designed to preserve performance while reducing FLOPs.

This paper presents AI-SNIPS (AI Support for Network Intelligence-based Pharmaceutical Security), a production-ready platform that enables stakeholder decision-making, secure data sharing, and interdisciplinary research in the fight against Illicit, Substandard, and Falsified Medical Products (ISFMP). AI-SNIPS takes as input cases: a case consists of one or more URLs suspected of ISFMP activity. Cases can be supplemented with ground-truth structured data (labeled keywords) such as seller PII or case notes. First, AI-SNIPS scrapes and stores relevant images and text from the provided URLs without any user intervention. Salient features for predicting case similarity are extracted from the aggregated data using a combination of rule-based and machine-learning techniques and used to construct a seller network, with the nodes representing cases (sellers) and the edges representing the similarity between two sellers. Network analysis and community detection techniques are applied to extract seller clusters ranked by profitability and their potential to harm society. Lastly, AI-SNIPS provides interpretability by distilling common word/image similarities for each cluster into signature vectors. We validate the importance of AI-SNIPS's features for distinguishing large pharmaceutical affiliate networks from small ISFMP operations using an actual ISFMP lead sheet.

As exemplified by the COVID-19 pandemic, our health and wellbeing depend on a difficult-to-measure web of societal factors and individual behaviors. This effort requires new algorithmic and data-driven paradigms which span the full process of gathering costly data, learning models to understand and predict such interactions, and optimizing the use of limited resources in interventions. In response to these needs, I present methodological developments at the intersection of machine learning, optimization, and social networks which are motivated by on-the-ground collaborations on HIV prevention, tuberculosis treatment, and the COVID-19 response. Here, I give an overview of two lines of work.

This work considers the use of AI and parallelism as a context for learning typical programming concepts in an introductory programming course (CS1). The course includes exercises in decision trees, a novel game called Find the Gnomes to introduce supervised learning, the construction and application of a vectorized neural network unit class, and obtaining speedup in training through parallelism. The exercises are designed to teach students typical introductory programming concepts while also providing a preview and motivating example of advanced CS topics. Students' understanding and motivation are considered through a detailed analysis of pre- and post-survey data gathered in several sections of the course each taught by one of four instructors across five semesters.

This demo paper discusses a scalable platform for emerging Data-Driven AI Applications targeted toward predictive maintenance solutions. We propose a common AI software architecture stack for building diverse AI Applications such as Anomaly Detection, Failure Pattern Analysis, Asset Health Forecasting, etc. for more than a 100K industrial assets of similar class. As a part of the AI system demonstration, we have identified the following three key topics for discussion: Scaling model training across multiple assets, Joint execution of multiple AI applications; and Bridge the gap between current open source software tools and the emerging need for AI Applications. To demonstrate the benefits, AI Model Factory has been tested to build the models for various industrial assets such as Wind turbines, Oil wells, etc. The system is deployed on API Hub for demonstration.

As artificial intelligence permeates our lives through various tools and services, there is an increasing need to consider how to teach young learners about AI in a relevant and engaging way. One way to do so is to leverage familiar and pervasive technologies such as conversational AIs. By learning about conversational AIs, learners are introduced to AI concepts such as computers’ perception of natural language, the need for training datasets, and the design of AI-human interactions. In this experience report, we describe a summer camp curriculum designed for middle school learners composed of general AI lessons, unplugged activities, conversational AI lessons, and project activities in which the campers develop their own conversational agents. The results show that this summer camp experience fostered significant increases in learners’ ability beliefs, willingness to share their learning experience, and intent to persist in AI learning. We conclude with a discussion of how conversational AI can be used as an entry point to K-12 AI education.

An essential element of K-12 AI literacy is educating learners about the ethical and societal implications of AI systems. Previous work in AI ethics literacy have developed curriculum and classroom activities that engage learners in reflecting on the ethical implications of AI systems and developing responsible AI. There is little work in using game-based learning methods in AI literacy. Games are known to be compelling media to teach children about complex STEM concepts. In this work, we developed a competitive card game for middle and high school students called “AI Audit” where they play as AI start-up founders building novel AI-powered technology. Players can challenge other players with potential harms of their technology or defend their own businesses by features that mitigate these harms. The game mechanics reward systems that are ethically developed or that take steps to mitigate potential harms. In this paper, we present the game design, teacher resources for classroom deployment and early playtesting results. We discuss our reflections about using games as teaching tools for AI literacy in K-12 classrooms.

The existing resource allocation policy for application instances in Kubernetes cannot dynamically adjust according to the requirement of business, which would cause an enormous waste of resources during fluctuations. Moreover, the emergence of new cloud services puts higher resource management requirements. This paper discusses horizontal POD resources management in Alibaba Cloud Container Services with a newly deployed AI algorithm framework named AHPA - the adaptive horizontal pod auto-scaling system. Based on a robust decomposition forecasting algorithm and performance training model, AHPA offers an optimal pod number adjustment plan that could reduce POD resources and maintain business stability. Since being deployed in April 2021, this system has expanded to multiple customer scenarios, including logistics, social networks, AI audio and video, e-commerce, etc. Compared with the previous algorithms, AHPA solves the elastic lag problem, increasing CPU usage by 10% and reducing resource cost by more than 20%. In addition, AHPA can automatically perform flexible planning according to the predicted business volume without manual intervention, significantly saving operation and maintenance costs.

Large-scale high-quality data is critical for training modern deep neural networks. However, data acquisition can be costly or time-consuming for many time-series applications, thus researchers turn to generative models for generating synthetic time-series data. In particular, recent generative adversarial networks (GANs) have achieved remarkable success in time-series generation. Despite their success, existing GAN models typically generate the sequences in an auto-regressive manner, and we empirically observe that they suffer from severe distribution shifts and bias amplification, especially when generating long sequences. To resolve this problem, we propose Adversarial Error Correction GAN (AEC-GAN), which is capable of dynamically correcting the bias in the past generated data to alleviate the risk of distribution shifts and thus can generate high-quality long sequences. AEC-GAN contains two main innovations: (1) We develop an error correction module to mitigate the bias. In the training phase, we adversarially perturb the realistic time-series data and then optimize this module to reconstruct the original data. In the generation phase, this module can act as an efficient regulator to detect and mitigate the bias. (2) We propose an augmentation method to facilitate GAN's training by introducing adversarial examples. Thus, AEC-GAN can generate high-quality sequences of arbitrary lengths, and the synthetic data can be readily applied to downstream tasks to boost their performance. We conduct extensive experiments on six widely used datasets and three state-of-the-art time-series forecasting models to evaluate the quality of our synthetic time-series data in different lengths and downstream tasks. Both the qualitative and quantitative experimental results demonstrate the superior performance of AEC-GAN over other deep generative models for time-series generation.

Existing works on anomaly detection (AD) rely on clean labels from human annotators that are expensive to acquire in practice. In this work, we propose a method to leverage weak/noisy labels (e.g., risk scores generated by machine rules for detecting malware) that are cheaper to obtain for anomaly detection. Specifically, we propose ADMoE, the first framework for anomaly detection algorithms to learn from noisy labels. In a nutshell, ADMoE leverages mixture-of-experts (MoE) architecture to encourage specialized and scalable learning from multiple noisy sources. It captures the similarities among noisy labels by sharing most model parameters, while encouraging specialization by building "expert" sub-networks. To further juice out the signals from noisy labels, ADMoE uses them as input features to facilitate expert learning. Extensive results on eight datasets (including a proprietary enterprise security dataset) demonstrate the effectiveness of ADMoE, where it brings up to 34% performance improvement over not using it. Also, it outperforms a total of 13 leading baselines with equivalent network parameters and FLOPS. Notably, ADMoE is model-agnostic to enable any neural network-based detection methods to handle noisy labels, where we showcase its results on both multiple-layer perceptron (MLP) and the leading AD method DeepSAD.

Several works have proven that finetuning is an applicable approach for debiasing contextualized word embeddings. Similarly, discrete prompts with semantic meanings have shown to be effective in debiasing tasks. With unfixed mathematical representation at the token level, continuous prompts usually surpass discrete ones at providing a pre-trained language model (PLM) with additional task-specific information. Despite this, relatively few efforts have been made to debias PLMs by prompt tuning with continuous prompts compared to its discrete counterpart. Furthermore, for most debiasing methods that alter a PLM's original parameters, a major problem is the need to not only decrease the bias in the PLM but also to ensure that the PLM does not lose its representation ability. Finetuning methods typically have a hard time maintaining this balance, as they tend to violently remove meanings of attribute words (like the words developing our concepts of "male" and "female" for gender), which also leads to an unstable and unpredictable training process. In this paper, we propose ADEPT, a method to debias PLMs using prompt tuning while maintaining the delicate balance between removing biases and ensuring representation ability. To achieve this, we propose a new training criterion inspired by manifold learning and equip it with an explicit debiasing term to optimize prompt tuning. In addition, we conduct several experiments with regard to the reliability, quality, and quantity of a previously proposed attribute training corpus in order to obtain a clearer prototype of a certain attribute, which indicates the attribute's position and relative distances to other words on the manifold. We evaluate ADEPT on several widely acknowledged debiasing benchmarks and downstream tasks, and find that it achieves competitive results while maintaining (and in some cases even improving) the PLM's representation ability. We further visualize words' correlation before and after debiasing a PLM, and give some possible explanations for the visible effects.

Automatic polyp segmentation from colonoscopy images is an essential prerequisite for the development of computer-assisted therapy. However, the complex semantic information and the blurred edges of polyps make segmentation extremely difficult. In this paper, we propose a novel semi-supervised polyp segmentation framework using affinity contrastive learning (ACL-Net), which is implemented between student and teacher networks to consistently refine the pseudo-labels for semi-supervised polyp segmentation. By aligning the affinity maps between the two branches, a better polyp region activation can be obtained to fully exploit the appearance-level context encoded in the feature maps, thereby improving the capability of capturing not only global localization and shape context, but also the local textural and boundary details. By utilizing the rich inter-image affinity context and establishing a global affinity context based on the memory bank, a cross-image affinity aggregation (CAA) module is also implemented to further refine the affinity aggregation between the two branches. By continuously and adaptively refining pseudo-labels with optimized affinity, we can improve the semi-supervised polyp segmentation based on the mutually reinforced knowledge interaction among contrastive learning and consistency learning iterations. Extensive experiments on five benchmark datasets, including Kvasir-SEG, CVC-ClinicDB, CVC-300, CVC-ColonDB and ETIS, demonstrate the effectiveness and superiority of our method. Codes are available at https://github.com/xiewende/ACL-Net.

Multi-agent reinforcement learning (MARL) suffers from the non-stationarity problem, which is the ever-changing targets at every iteration when multiple agents update their policies at the same time. Starting from first principle, in this paper, we manage to solve the non-stationarity problem by proposing bidirectional action-dependent Q-learning (ACE). Central to the development of ACE is the sequential decision making process wherein only one agent is allowed to take action at one time. Within this process, each agent maximizes its value function given the actions taken by the preceding agents at the inference stage. In the learning phase, each agent minimizes the TD error that is dependent on how the subsequent agents have reacted to their chosen action. Given the design of bidirectional dependency, ACE effectively turns a multi-agent MDP into a single-agent MDP. We implement the ACE framework by identifying the proper network representation to formulate the action dependency, so that the sequential decision process is computed implicitly in one forward pass. To validate ACE, we compare it with strong baselines on two MARL benchmarks. Empirical experiments demonstrate that ACE outperforms the state-of-the-art algorithms on Google Research Football and StarCraft Multi-Agent Challenge by a large margin. In particular, on SMAC tasks, ACE achieves 100% success rate on almost all the hard and super hard maps. We further study extensive research problems regarding ACE, including extension, generalization and practicability.