IJCAI 2022 论文列表

Stock recommendation plays a critical role in modern quantitative trading. The large volumes of social media information such as investment reviews that delegate emotion-driven factors, together with price technical indicators formulate a “snapshot” of the evolving stock market profile. However, previous studies usually model the temporal trajectories of price and media modalities separately while losing their interrelated influences. Moreover, they mainly extract review semantics via sequential or attentive models, whereas the rich text associated knowledge is largely neglected. In this paper, we propose a novel heterogeneous interactive snapshot network for stock profiling and recommendation. We model investment reviews in each snapshot as a heterogeneous document graph, and develop a flexible hierarchical attentive propagation framework to capture fine-grained proximity features. Further, to learn stock embedding for ranking, we introduce a novel twins-GRU method, which tightly couples the media and price parallel sequences in a cross-interactive fashion to catch dynamic dependencies between successive snapshots. Our approach excels state-of-the-arts over 7.6% in terms of cumulative and risk-adjusted returns in trading simulations on both English and Chinese benchmarks.

Federated learning (FL) enables edge-devices to collaboratively learn a model without disclosing their private data to a central aggregating server. Most existing FL algorithms require models of identical architecture to be deployed across the clients and server, making it infeasible to train large models due to clients' limited system resources. In this work, we propose a novel ensemble knowledge transfer method named Fed-ET in which small models (different in architecture) are trained on clients, and used to train a larger model at the server. Unlike in conventional ensemble learning, in FL the ensemble can be trained on clients' highly heterogeneous data. Cognizant of this property, Fed-ET uses a weighted consensus distillation scheme with diversity regularization that efficiently extracts reliable consensus from the ensemble while improving generalization by exploiting the diversity within the ensemble. We show the generalization bound for the ensemble of weighted models trained on heterogeneous datasets that supports the intuition of Fed-ET. Our experiments on image and language tasks show that Fed-ET significantly outperforms other state-of-the-art FL algorithms with fewer communicated parameters, and is also robust against high data-heterogeneity.

Data sets composed of a mixture of categorical and numerical attributes (also called mixed data hereinafter) are common in real-world cluster analysis. However, insightful analysis of such data under an unsupervised scenario using clustering is extremely challenging because the information provided by the two different types of attributes is heterogeneous, being at different concept hierarchies. That is, the values of a categorical attribute represent a set of different concepts (e.g., professor, lawyer, and doctor of the attribute "occupation"), while the values of a numerical attribute describe the tendencies toward two different concepts (e.g., low and high of the attribute "income"). To appropriately use such heterogeneous information in clustering, this paper therefore proposes a novel attribute representation learning method called Het2Hom, which first converts the heterogeneous attributes into a homogeneous form, and then learns attribute representations and data partitions on such a homogeneous basis. Het2Hom features low time complexity and intuitive interpretability. Extensive experiments show that Het2Hom outperforms the state-of-the-art counterparts.

Sequences of group interactions, such as emails, online discussions, and co-authorships, are ubiquitous; and they are naturally represented as a stream of hyperedges (i.e., sets of nodes). Despite its broad potential applications, anomaly detection in hypergraphs (i.e., sets of hyperedges) has received surprisingly little attention, compared to anomaly detection in graphs. While it is tempting to reduce hypergraphs to graphs and apply existing graph-based methods, according to our experiments, taking higher-order structures of hypergraphs into consideration is worthwhile. We propose HashNWalk, an incremental algorithm that detects anomalies in a stream of hyperedges. It maintains and updates a constant-size summary of the structural and temporal information about the input stream. Using the summary, which is the form of a proximity matrix, HashNWalk measures the anomalousness of each new hyperedge as it appears. HashNWalk is (a) Fast: it processes each hyperedge in near real-time and billions of hyperedges within a few hours, (b) Space Efficient: the size of the maintained summary is a user-specific constant, (c) Effective: it successfully detects anomalous hyperedges in real-world hypergraphs.

With the rise in popularity of short-term Human Trajectory Prediction (HTP), Long-Term Crowd Flow Prediction (LTCFP) has been proposed to forecast crowd movement in large and complex environments. However, the input representations, models, and datasets for LTCFP are currently limited. To this end, we propose Fourier Isovists, a novel input representation based on egocentric visibility, which consistently improves all existing models. We also propose GeoInteractNet (GINet), which couples the layers between a multi-scale attention network (M-SCAN) and a convolutional encoder-decoder network (CED). M-SCAN approximates a super-resolution map of where humans are likely to interact on the way to their goals and produces multi-scale attention maps. The CED then uses these maps in either its encoder's inputs or its decoder's attention gates, which allows GINet to produce super-resolution predictions with substantially higher accuracy than existing models even with Fourier Isovists. In order to evaluate the scalability of models to large and complex environments, which the only existing LTCFP dataset is unsuitable for, a new synthetic crowd dataset with both real and synthetic environments has been generated. In its nascent state, LTCFP has much to gain from our key contributions. The Supplementary Materials, dataset, and code are available at sssohn.github.io/GeoInteractNet.

The diversified top-k weight clique (DTKWC) search problem is an important generalization of the diversified top-k clique (DTKC) search problem with extensive applications, which extends the DTKC search problem by taking into account the weight of vertices. In this paper, we formulate DTKWC search problem using mixed integer linear program constraints and propose an efficient hybrid evolutionary algorithm (HEA-D) that combines a clique-based crossover operator and an effective simulated annealing-based local optimization procedure to find high-quality local optima. The experimental results show that HEA-D performs much better than the existing methods on two representative real-world benchmarks.

The ever-increasing data scale of user-item interactions makes it challenging for an effective and efficient recommender system. Recently, hash-based collaborative filtering (Hash-CF) approaches employ efficient Hamming distance of learned binary representations of users and items to accelerate recommendations. However, Hash-CF often faces two challenging problems, i.e., optimization on discrete representations and preserving semantic information in learned representations. To address the above two challenges, we propose HCFRec, a novel Hash-CF approach for effective and efficient recommendations. Specifically, HCFRec not only innovatively introduces normalized flow to learn the optimal hash code by efficiently fitting a proposed approximate mixture multivariate normal distribution, a continuous but approximately discrete distribution, but also deploys a cluster consistency preserving mechanism to preserve the semantic structure in representations for more accurate recommendations. Extensive experiments conducted on six real-world datasets demonstrate the superiority of our HCFRec compared to the state-of-art methods in terms of effectiveness and efficiency.

We study epistemic voting on binary issues where voters are characterized by their competence, i.e., the probability of voting for the correct alternative, and can choose between two actions: voting or abstaining. In our setting voting involves the expenditure of some effort, which is required to achieve the appropriate level of competence, whereas abstention carries no effort. We model this scenario as a game and characterize its equilibria under several variations. Our results show that when agents are aware of everyone's incentives, then the addition of effort may lead to Nash equilibria where wisdom of the crowds is lost. We further show that if agents' awareness of each other is constrained by a social network, the topology of the network may actually mitigate this effect.

Learning the similarity between structured data, especially the graphs, is one of the essential problems. Besides the approach like graph kernels, Gromov-Wasserstein (GW) distance recently draws a big attention due to its flexibility to capture both topological and feature characteristics, as well as handling the permutation invariance. However, structured data are widely distributed for different data mining and machine learning applications. With privacy concerns, accessing the decentralized data is limited to either individual clients or different silos. To tackle these issues, we propose a privacy-preserving framework to analyze the GW discrepancy of node embedding learned locally from graph neural networks in a federated flavor, and then explicitly place local differential privacy (LDP) based on Multi-bit Encoder to protect sensitive information. Our experiments show that, with strong privacy protection guaranteed by ε-LDP algorithm, the proposed framework not only preserves privacy in graph learning, but also presents a noised structural metric under GW distance, resulting in comparable and even better performance in classification and clustering tasks. Moreover, we reason the rationale behind the LDP-based GW distance analytically and empirically.

Graph classification is a challenging research task in many applications across a broad range of domains. Recently, Graph Neural Network (GNN) models have achieved superior performance on various real-world graph datasets. Despite their successes, most of current GNN models largely suffer from the ubiquitous class imbalance problem, which typically results in prediction bias towards majority classes. Although many imbalanced learning methods have been proposed, they mainly focus on regular Euclidean data and cannot well utilize topological structure of graph (non-Euclidean) data. To boost the performance of GNNs and investigate the relationship between topological structure and class imbalance, we propose GraphDIVE, which learns multi-view graph representations and combine multi-view experts (i.e., classifiers). Specifically, multi-view graph representations correspond to the intrinsic diverse graph topological structure characteristics. Extensive experiments on molecular benchmark datasets demonstrate the effectiveness of the proposed approach.

As time goes by, language evolves with word semantics changing. Unfortunately, traditional word embedding methods neglect the evolution of language and assume that word representations are static. Although contextualized word embedding models can capture the diverse representations of polysemous words, they ignore temporal information as well. To tackle the aforementioned challenges, we propose a graph-based dynamic word embedding (GDWE) model, which focuses on capturing the semantic drift of words continually. We introduce word-level knowledge graphs (WKGs) to store short-term and long-term knowledge. WKGs can provide rich structural information as supplement of lexical information, which help enhance the word embedding quality and capture semantic drift quickly. Theoretical analysis and extensive experiments validate the effectiveness of our GDWE on dynamic word embedding learning.

Performance estimation of neural architecture is a crucial component of neural architecture search (NAS). Meanwhile, neural predictor is a current mainstream performance estimation method. However, it is a challenging task to train the predictor with few architecture evaluations for efficient NAS. In this paper, we propose a graph masked autoencoder (GMAE) enhanced predictor, which can reduce the dependence on supervision data by self-supervised pre-training with untrained architectures. We compare our GMAE-enhanced predictor with existing predictors in different search spaces, and experimental results show that our predictor has high query utilization. Moreover, GMAE-enhanced predictor with different search strategies can discover competitive architectures in different search spaces. Code and supplementary materials are available at https://github.com/kunjing96/GMAENAS.git.

Word embedding has been widely used in various areas to boost the performance of the neural models. However, when processing context-free languages, embedding grammar rules with word embedding loses two types of information. One is the structural relationship between the grammar rules, and the other one is the content information of the rule definition. In this paper, we make the first attempt to learn a grammar-preserving rule embedding. We first introduce a novel graph structure to represent the context-free grammar. Then, we apply a Graph Neural Network (GNN) to extract the structural information and use a gating layer to integrate content information. We conducted experiments on six widely-used benchmarks containing four context-free languages. The results show that our approach improves the accuracy of the base model by 0.8 to 6.4 percentage points. Furthermore, Grape also achieves 1.6 F1 score improvement on the method naming task which shows the generality of our approach.

Goal-conditioned reinforcement learning (GCRL), related to a set of complex RL problems, trains an agent to achieve different goals under particular scenarios. Compared to the standard RL solutions that learn a policy solely depending on the states or observations, GCRL additionally requires the agent to make decisions according to different goals. In this survey, we provide a comprehensive overview of the challenges and algorithms for GCRL. Firstly, we answer what the basic problems are studied in this field. Then, we explain how goals are represented and present how existing solutions are designed from different points of view. Finally, we make the conclusion and discuss potential future prospects that recent researches focus on.

Although multistage tasks involving multiple sequential goals are common in real-world applications, they are not fully studied in multi-agent reinforcement learning (MARL). To accomplish a multi-stage task, agents have to achieve cooperation on different subtasks. Exploring the collaborative patterns of different subtasks and the sequence of completing the subtasks leads to an explosion in the search space, which poses great challenges to policy learning. Existing works designed for single-stage tasks where agents learn to cooperate only once usually suffer from low sample efficiency in multi-stage tasks as agents explore aimlessly. Inspired by human’s improving cooperation through goal consistency, we propose Multi-Agent Goal Consistency (MAGIC) framework to improve sample efficiency for learning in multi-stage tasks. MAGIC adopts a goal-oriented actor-critic model to learn both local and global views of goal cognition, which helps agents understand the task at the goal level so that they can conduct targeted exploration accordingly. Moreover, to improve exploration efficiency, MAGIC employs two-level goal consistency training to drive agents to formulate a consistent goal cognition. Experimental results show that MAGIC significantly improves sample efficiency and facilitates cooperation among agents compared with state-of-art MARL algorithms in several challenging multistage tasks.

Recent research attention for relation extraction has been paid to the dialogue scenario, i.e., dialogue-level relation extraction (DiaRE). Existing DiaRE methods either simply concatenate the utterances in a dialogue into a long piece of text, or employ naive words, sentences or entities to build dialogue graphs, while the structural characteristics in dialogues have not been fully utilized. In this work, we investigate a novel dialogue-level mixed dependency graph (D2G) and an argument reasoning graph (ARG) for DiaRE with a global relation reasoning mechanism. First, we model the entire dialogue into a unified and coherent D2G by explicitly integrating both syntactic and discourse structures, which enables richer semantic and feature learning for relation extraction. Second, we stack an ARG graph on top of D2G to further focus on argument inter-dependency learning and argument representation refinement, for sufficient argument relation inference. In our global reasoning framework, D2G and ARG work collaboratively, iteratively performing lexical, syntactic and semantic information exchange and representation learning over the entire dialogue context. On two DiaRE benchmarks, our framework shows considerable improvements over the current state-of-the-art baselines. Further analyses show that the model effectively solves the long-range dependence issue, and meanwhile gives explainable predictions.

With the increasing popularity of food delivery platforms, it has become pertinent to look into the working conditions of the `gig' workers in these platforms, especially providing them fair wages, reasonable working hours, and transparency on work availability. However, any solution to these problems must not degrade customer experience and be cost-effective to ensure that platforms are willing to adopt them. We propose Work4Food, which provides income guarantees to delivery agents, while minimizing platform costs and ensuring customer satisfaction. Work4Food ensures that the income guarantees are met in such a way that it does not lead to increased working hours or degrade environmental impact. To incorporate these objectives, Work4Food balances supply and demand by controlling the number of agents in the system and providing dynamic payment guarantees to agents based on factors such as agent location, ratings, etc. We evaluate Work4Food on a real-world dataset from a leading food delivery platform and establish its advantages over the state of the art in terms of the multi-dimensional objectives at hand.

Transformers have become methods of choice in many applications thanks to their ability to represent complex interactions between elements. However, extending the Transformer architecture to non-sequential data such as molecules and enabling its training on small datasets remains a challenge. In this work, we introduce a Transformer-based architecture for molecule property prediction, which is able to capture the geometry of the molecule. We modify the classical positional encoder by an initial encoding of the molecule geometry, as well as a learned gated self-attention mechanism. We further suggest an augmentation scheme for molecular data capable of avoiding the overfitting induced by the overparameterized architecture. The proposed framework outperforms the state-of-the-art methods while being based on pure machine learning solely, i.e. the method does not incorporate domain knowledge from quantum chemistry and does not use extended geometric inputs besides the pairwise atomic distances.

Writing a survey paper on one research topic usually needs to cover the salient content from numerous related papers, which can be modeled as a multi-document summarization (MDS) task. Existing MDS datasets usually focus on producing the structureless summary covering a few input documents. Meanwhile, previous structured summary generation works focus on summarizing a single document into a multi-section summary. These existing datasets and methods cannot meet the requirements of summarizing numerous academic papers into a structured summary. To deal with the scarcity of available data, we propose BigSurvey, the first large-scale dataset for generating comprehensive summaries of numerous academic papers on each topic. We collect target summaries from more than seven thousand survey papers and utilize their 430 thousand reference papers’ abstracts as input documents. To organize the diverse content from dozens of input documents and ensure the efficiency of processing long text sequences, we propose a summarization method named category-based alignment and sparse transformer (CAST). The experimental results show that our CAST method outperforms various advanced summarization methods.