AAAI 2023 论文列表

Learning exercises that activate students’ additional cognitive understanding of course concepts facilitate contextualizing the content knowledge and developing higher-order thinking and problem-solving skills. Student-generated instructional materials such as course summaries and problem sets are amongst the instructional strategies that reflect active learning and constructivist philosophy. The contributions of this work are twofold: 1) We introduce a practical implementation of inside-outside learning strategy in an undergraduate deep learning course and will share our experiences in incorporating student-generated instructional materials learning strategy in course design, and 2) We develop a context-aware deep learning framework to draw insights from the student-generated materials for (i) Detecting anomalies in group activities and (ii) Predicting the median quiz performance of students in each group. This work opens up an avenue for effectively implementing a constructivism learning strategy in large-scale and online courses to build a sense of community between learners while providing an automated tool for instructors to identify at-risk groups.

Interpretations of logical formulas over semirings (other than the Boolean semiring) have applications in various areas of computer science including logic, AI, databases, and security. Such interpretations provide richer information beyond the truth or falsity of a statement. Examples of such semirings include Viterbi semiring, min-max or access control semiring, tropical semiring, and fuzzy semiring. The present work investigates the complexity of constraint optimization problems over semirings. The generic optimization problem we study is the following: Given a propositional formula phi over n variable and a semiring (K,+, . ,0,1), find the maximum value over all possible interpretations of phi over K. This can be seen as a generalization of the well-known satisfiability problem (a propositional formula is satisfiable if and only if the maximum value over all interpretations/assignments over the Boolean semiring is 1). A related problem is to find an interpretation that achieves the maximum value. In this work, we first focus on these optimization problems over the Viterbi semiring, which we call optConfVal and optConf. We first show that for general propositional formulas in negation normal form, optConfVal and optConf are in FP^NP. We then investigate optConf when the input formula phi is represented in the conjunctive normal form. For CNF formulae, we first derive an upper bound on the value of optConf as a function of the number of maximum satisfiable clauses. In particular, we show that if r is the maximum number of satisfiable clauses in a CNF formula with m clauses, then its optConf value is at most 1/4^(m-r). Building on this we establish that optConf for CNF formulae is hard for the complexity class FP^NP[log]. We also design polynomial-time approximation algorithms and establish an inapproximability for optConfVal. We establish similar complexity results for these optimization problems over other semirings including tropical, fuzzy, and access control semirings.

Advances in machine learning have enabled the prediction of immune system responses to prophylactic and therapeutic vaccines. However, the engineering task of designing vaccines remains a challenge. In particular, the genetic variability of the human immune system makes it difficult to design peptide vaccines that provide widespread immunity in vaccinated populations. We introduce a framework for evaluating and designing peptide vaccines that uses probabilistic machine learning models, and demonstrate its ability to produce designs for a SARS-CoV-2 vaccine that outperform previous designs. We provide a theoretical analysis of the approximability, scalability, and complexity of our framework.

One approach to guaranteeing safety in Reinforcement Learning is through cost constraints that are dependent on the policy. Recent works in constrained RL have developed methods that ensure constraints are enforced even at learning time while maximizing the overall value of the policy. Unfortunately, as demonstrated in our experimental results, such approaches do not perform well on complex multi-level tasks, with longer episode lengths or sparse rewards. To that end, we propose a scalable hierarchical approach for constrained RL problems that employs backward cost value functions in the context of task hierarchy and a novel intrinsic reward function in lower levels of the hierarchy to enable cost constraint enforcement. One of our key contributions is in proving that backward value functions are theoretically viable even when there are multiple levels of decision making. We also show that our new approach, referred to as Hierarchically Limited consTraint Enforcement (HiLiTE) significantly improves on state of the art Constrained RL approaches for many benchmark problems from literature. We further demonstrate that this performance (on value and constraint enforcement) clearly outperforms existing best approaches for constrained RL and hierarchical RL.

With the rapid development of the airline industry, maximizing the market share with a constrained budget is an urgent econometric problem for an airline. We investigate the problem by adjusting flight frequencies on different flight routes. Owing to the large search space of solutions and the difficulty of predicting the market, this problem is in general daunting to solve. This paper proposes a novel two-stage optimization method to address the challenges. On the higher level, we use a signal to guide the optimization process toward a constrained satisfying solution. On the lower level, we consider the consecutive itineraries in real scenarios and model the unseen correlations between routes in itineraries for market share prediction. In theory, we prove the convergence of our optimization approach. In the experiment, we empirically verify the superiority of both our prediction model and optimization approach over existing works with large-scale real-world data. Our code has been released at: https://github.com/codingAndBS/AirlineMarket.

Almost all multi-agent reinforcement learning algorithms without communication follow the principle of centralized training with decentralized execution. During the centralized training, agents can be guided by the same signals, such as the global state. However, agents lack the shared signal and choose actions given local observations during execution. Inspired by viewpoint invariance and contrastive learning, we propose consensus learning for cooperative multi-agent reinforcement learning in this study. Although based on local observations, different agents can infer the same consensus in discrete spaces without communication. We feed the inferred one-hot consensus to the network of agents as an explicit input in a decentralized way, thereby fostering their cooperative spirit. With minor model modifications, our suggested framework can be extended to a variety of multi-agent reinforcement learning algorithms. Moreover, we carry out these variants on some fully cooperative tasks and get convincing results.

Nowadays, systems based on machine learning (ML) are widely used in different domains. Given their popularity, ML models have become targets for various attacks. As a result, research at the intersection of security/privacy and ML has flourished. Typically such work has focused on individual types of security/privacy concerns and mitigations thereof. However, in real-life deployments, an ML model will need to be protected against several concerns simultaneously. A protection mechanism optimal for a specific security or privacy concern may interact negatively with mechanisms intended to address other concerns. Despite its practical relevance, the potential for such conflicts has not been studied adequately. In this work, we first provide a framework for analyzing such conflicting interactions. We then focus on systematically analyzing pairwise interactions between protection mechanisms for one concern, model and data ownership verification, with two other classes of ML protection mechanisms: differentially private training, and robustness against model evasion. We find that several pairwise interactions result in conflicts. We also explore potential approaches for avoiding such conflicts. First, we study the effect of hyperparameter relaxations, finding that there is no sweet spot balancing the performance of both protection mechanisms. Second, we explore whether modifying one type of protection mechanism (ownership verification) so as to decouple it from factors that may be impacted by a conflicting mechanism (differentially private training or robustness to model evasion) can avoid conflict. We show that this approach can indeed avoid the conflict between ownership verification mechanisms when combined with differentially private training, but has no effect on robustness to model evasion. We conclude by identifying the gaps in the landscape of studying interactions between other types of ML protection mechanisms.

Any classifier can be "smoothed out" under Gaussian noise to build a new classifier that is provably robust to l2-adversarial perturbations, viz., by averaging its predictions over the noise via randomized smoothing. Under the smoothed classifiers, the fundamental trade-off between accuracy and (adversarial) robustness has been well evidenced in the literature: i.e., increasing the robustness of a classifier for an input can be at the expense of decreased accuracy for some other inputs. In this paper, we propose a simple training method leveraging this trade-off to obtain robust smoothed classifiers, in particular, through a sample-wise control of robustness over the training samples. We make this control feasible by using "accuracy under Gaussian noise" as an easy-to-compute proxy of adversarial robustness for an input. Specifically, we differentiate the training objective depending on this proxy to filter out samples that are unlikely to benefit from the worst-case (adversarial) objective. Our experiments show that the proposed method, despite its simplicity, consistently exhibits improved certified robustness upon state-of-the-art training methods. Somewhat surprisingly, we find these improvements persist even for other notions of robustness, e.g., to various types of common corruptions. Code is available at https://github.com/alinlab/smoothing-catrs.

Syntax splitting is a property of inductive inference operators that ensures we can restrict our attention to parts of the conditional belief base that share atoms with a given query. To apply syntax splitting, a conditional belief base needs to consist of syntactically disjoint conditionals. This requirement is often too strong in practice, as conditionals might share atoms. In this paper we introduce the concept of conditional syntax splitting, inspired by the notion of conditional independence as known from probability theory. We show that lexicographic inference and system W satisfy conditional syntax splitting, and connect conditional syntax splitting to several known properties from the literature on non-monotonic reasoning, including the drowning effect.

Learning the underlying distribution of molecular graphs and generating high-fidelity samples is a fundamental research problem in drug discovery and material science. However, accurately modeling distribution and rapidly generating novel molecular graphs remain crucial and challenging goals. To accomplish these goals, we propose a novel Conditional Diffusion model based on discrete Graph Structures (CDGS) for molecular graph generation. Specifically, we construct a forward graph diffusion process on both graph structures and inherent features through stochastic differential equations (SDE) and derive discrete graph structures as the condition for reverse generative processes. We present a specialized hybrid graph noise prediction model that extracts the global context and the local node-edge dependency from intermediate graph states. We further utilize ordinary differential equation (ODE) solvers for efficient graph sampling, based on the semi-linear structure of the probability flow ODE. We also combine the solvers with gradient guidance from the molecule property predictor for similarity-constrained molecule optimization. Experiments on diverse datasets validate the effectiveness of our framework. Particularly, the proposed method still generates high-quality molecular graphs in a limited number of steps.

Streams of irregularly occurring events are commonly modeled as a marked temporal point process. Many real-world datasets such as e-commerce transactions and electronic health records often involve events where multiple event types co-occur, e.g. multiple items purchased or multiple diseases diagnosed simultaneously. In this paper, we tackle multi-label prediction in such a problem setting, and propose a novel Transformer-based Conditional Mixture of Bernoulli Network (TCMBN) that leverages neural density estimation to capture complex temporal dependence as well as probabilistic dependence between concurrent event types. We also propose potentially incorporating domain knowledge in the objective by regularizing the predicted probability. To represent probabilistic dependence of concurrent event types graphically, we design a two-step approach that first learns the mixture of Bernoulli network and then solves a least-squares semi-definite constrained program to numerically approximate the sparse precision matrix from a learned covariance matrix. This approach proves to be effective for event prediction while also providing an interpretable and possibly non-stationary structure for insights into event co-occurrence. We demonstrate the superior performance of our approach compared to existing baselines on multiple synthetic and real benchmarks.

Despite the broad application of deep reinforcement learning (RL), transferring and adapting the policy to unseen but similar environments is still a significant challenge. Recently, the language-conditioned policy is proposed to facilitate policy transfer through learning the joint representation of observation and text that catches the compact and invariant information across various environments. Existing studies of language-conditioned RL methods often learn the joint representation as a simple latent layer for the given instances (episode-specific observation and text), which inevitably includes noisy or irrelevant information and cause spurious correlations that are dependent on instances, thus hurting generalization performance and training efficiency. To address the above issue, we propose a conceptual reinforcement learning (CRL) framework to learn the concept-like joint representation for language-conditioned policy. The key insight is that concepts are compact and invariant representations in human cognition through extracting similarities from numerous instances in real-world. In CRL, we propose a multi-level attention encoder and two mutual information constraints for learning compact and invariant concepts. Verified in two challenging environments, RTFM and Messenger, CRL significantly improves the training efficiency (up to 70%) and generalization ability (up to 30%) to the new environment dynamics.

The opacity of deep neural networks remains a challenge in deploying solutions where explanation is as important as precision. We present ConceptX, a human-in-the-loop framework for interpreting and annotating latent representational space in pre-trained Language Models (pLMs). We use an unsupervised method to discover concepts learned in these models and enable a graphical interface for humans to generate explanations for the concepts. To facilitate the process, we provide auto-annotations of the concepts (based on traditional linguistic ontologies). Such annotations enable development of a linguistic resource that directly represents latent concepts learned within deep NLP models. These include not just traditional linguistic concepts, but also task-specific or sensitive concepts (words grouped based on gender or religious connotation) that helps the annotators to mark bias in the model. The framework consists of two parts (i) concept discovery and (ii) annotation platform.

Dense passage retrieval aims to retrieve the relevant passages of a query from a large corpus based on dense representations (i.e., vectors) of the query and the passages. Recent studies have explored improving pre-trained language models to boost dense retrieval performance. This paper proposes CoT-MAE (ConTextual Masked Auto-Encoder), a simple yet effective generative pre-training method for dense passage retrieval. CoT-MAE employs an asymmetric encoder-decoder architecture that learns to compress the sentence semantics into a dense vector through self-supervised and context-supervised masked auto-encoding. Precisely, self-supervised masked auto-encoding learns to model the semantics of the tokens inside a text span, and context-supervised masked auto-encoding learns to model the semantical correlation between the text spans. We conduct experiments on large-scale passage retrieval benchmarks and show considerable improvements over strong baselines, demonstrating the high efficiency of CoT-MAE. Our code is available at https://github.com/caskcsg/ir/tree/main/cotmae.

There are many applications that benefit from computing the exact divergence between 2 discrete probability measures, including machine learning. Unfortunately, in the absence of any assumptions on the structure or independencies within these distributions, computing the divergence between them is an intractable problem in high dimensions. We show that we are able to compute a wide family of functionals and divergences, such as the alpha-beta divergence, between two decomposable models, i.e. chordal Markov networks, in time exponential to the treewidth of these models. The alpha-beta divergence is a family of divergences that include popular divergences such as the Kullback-Leibler divergence, the Hellinger distance, and the chi-squared divergence. Thus, we can accurately compute the exact values of any of this broad class of divergences to the extent to which we can accurately model the two distributions using decomposable models.

In reinforcement learning, the classic objectives of maximizing discounted and finite-horizon cumulative rewards are PAC-learnable: There are algorithms that learn a near-optimal policy with high probability using a finite amount of samples and computation. In recent years, researchers have introduced objectives and corresponding reinforcement-learning algorithms beyond the classic cumulative rewards, such as objectives specified as linear temporal logic formulas. However, questions about the PAC-learnability of these new objectives have remained open. This work demonstrates the PAC-learnability of general reinforcement-learning objectives through sufficient conditions for PAC-learnability in two analysis settings. In particular, for the analysis that considers only sample complexity, we prove that if an objective given as an oracle is uniformly continuous, then it is PAC-learnable. Further, for the analysis that considers computational complexity, we prove that if an objective is computable, then it is PAC-learnable. In other words, if a procedure computes successive approximations of the objective's value, then the objective is PAC-learnable. We give three applications of our condition on objectives from the literature with previously unknown PAC-learnability and prove that these objectives are PAC-learnable. Overall, our result helps verify existing objectives' PAC-learnability. Also, as some studied objectives that are not uniformly continuous have been shown to be not PAC-learnable, our results could guide the design of new PAC-learnable objectives.

Transformer and its variants achieve excellent results in various computer vision and natural language processing tasks, but high computational costs and reliance on large training datasets restrict their deployment in resource-constrained settings. Low-rank approximation of model weights has been effective in compressing CNN models, but its application to transformers has been less explored and is less effective. Existing methods require the complete dataset to fine-tune compressed models, which are both time-consuming and data-hungry. This paper reveals that the features (i.e., activations) are low-rank, but model weights are surprisingly not low-rank. Hence, AAFM is proposed, which adaptively determines the compressed model structure and locally compresses each linear layer's output features rather than the model weights. A second stage, GFM, optimizes the entire compressed network holistically. Both AAFM and GFM only use few training samples without labels, that is, they are few-shot, unsupervised, fast and effective. For example, with only 2K images without labels, 33% of the parameters are removed in DeiT-B with 18.8% relative throughput increase, but only a 0.23% accuracy loss for ImageNet recognition. The proposed methods are successfully applied to the language modeling task in NLP, too. Besides, the few-shot compressed models generalize well in downstream tasks.

Experience management is an emerging business area where organizations focus on understanding the feedback of customers and employees in order to improve their end-to-end experiences. This results in a unique set of machine learning problems to help understand how people feel, discover issues they care about, and find which actions need to be taken on data that are different in content and distribution from traditional NLP domains. In this paper, we present a case study of building text analysis applications that perform multiple classification tasks efficiently in 12 languages in the nascent business area of experience management. In order to scale up modern ML methods on experience data, we leverage cross lingual and multi-task modeling techniques to consolidate our models into a single deployment to avoid overhead. We also make use of model compression and model distillation to reduce overall inference latency and hardware cost to the level acceptable for business needs while maintaining model prediction quality. Our findings show that multi-task modeling improves task performance for a subset of experience management tasks in both XLM-R and mBert architectures. Among the compressed architectures we explored, we found that MiniLM achieved the best compression/performance tradeoff. Our case study demonstrates a speedup of up to 15.61x with 2.60% average task degradation (or 3.29x speedup with 1.71% degradation) and estimated savings of 44% over using the original full-size model. These results demonstrate a successful scaling up of text classification for the challenging new area of ML for experience management.

Multi-document summarization (MDS) aims to generate a summary for a number of related documents. We propose HGSum — an MDS model that extends an encoder-decoder architecture to incorporate a heterogeneous graph to represent different semantic units (e.g., words and sentences) of the documents. This contrasts with existing MDS models which do not consider different edge types of graphs and as such do not capture the diversity of relationships in the documents. To preserve only key information and relationships of the documents in the heterogeneous graph, HGSum uses graph pooling to compress the input graph. And to guide HGSum to learn the compression, we introduce an additional objective that maximizes the similarity between the compressed graph and the graph constructed from the ground-truth summary during training. HGSum is trained end-to-end with the graph similarity and standard cross-entropy objectives. Experimental results over Multi-News, WCEP-100, and Arxiv show that HGSum outperforms state-of-the-art MDS models. The code for our model and experiments is available at: https://github.com/oaimli/HGSum.