Online recruitment platforms have reshaped job-seeking and recruiting processes, driving increased demand for applications that enhance person-job matching. Traditional methods generally rely on analyzing textual data from resumes and job descriptions, limiting the dynamic, interactive aspects crucial to effective recruitment. Recent advances in Large Language Models (LLMs) have revealed remarkable potential in simulating adaptive, role-based dialogues, making them well-suited for recruitment scenarios. In this paper, we propose MockLLM, a novel framework to generate and evaluate mock interview interactions. The system consists of two key components: mock interview generation and two-sided evaluation in handshake protocol. By simulating both interviewer and candidate roles, MockLLM enables consistent and collaborative interactions for real-time and two-sided matching. To further improve the matching quality, MockLLM further incorporates reflection memory generation and dynamic strategy modification, refining behaviors based on previous experience. We evaluate MockLLM on real-world data Boss Zhipin, a major Chinese recruitment platform. The experimental results indicate that MockLLM outperforms existing methods in matching accuracy, scalability, and adaptability across job domains, highlighting its potential to advance candidate assessment and online recruitment.

Evaluation of summary quality is particularly crucial within the scientific domain, because it facilitates efficient knowledge dissemination and automated scientific information retrieval. This article presents conceptual and experimental analyses of scientific summarization, highlighting the inadequacies of traditional evaluation methods. These methods, including \( n \) -gram overlap calculations, embedding comparisons, verification, and QA-based approaches, often fall short in providing explanations, grasping scientific concepts, or identifying key content. Correspondingly, we introduce the Facet-aware Metric (FM), employing LLMs for advanced semantic matching to evaluate summaries based on different facets. The facet granularity is tailored to the structure of scientific abstracts, offering an integrated evaluation approach that is not fragmented, while also providing fine-grained interpretability. Recognizing the absence of an evaluation benchmark in the scientific domain, we curate a Scientific abstract summary evaluation Dataset (ScholarSum) with facet-level annotations. Our findings confirm that FM offers a more logical approach to evaluating scientific summaries. In addition, fine-tuned smaller models can compete with LLMs in scientific contexts, while LLMs have limitations in learning from in-context information in scientific domains. We hope our benchmark inspires better evaluation metrics and future enhancements to LLMs: https://github.com/iriscxy/ScholarSum .

A proficient summarization model should exhibit both flexibility -- the capacity to handle a range of in-domain summarization tasks, and adaptability -- the competence to acquire new knowledge and adjust to unseen out-of-domain tasks. Unlike large language models (LLMs) that achieve this through parameter scaling, we propose a more parameter-efficient approach in this study. Our motivation rests on the principle that the general summarization ability to capture salient information can be shared across different tasks, while the domain-specific summarization abilities need to be distinct and tailored. Concretely, we propose MoeSumm, a Mixture-of-Expert Summarization architecture, which utilizes a main expert for gaining the general summarization capability and deputy experts that selectively collaborate to meet specific summarization task requirements. We further propose a max-margin loss to stimulate the separation of these abilities. Our model's distinct separation of general and domain-specific summarization abilities grants it with notable flexibility and adaptability, all while maintaining parameter efficiency. MoeSumm achieves flexibility by managing summarization across multiple domains with a single model, utilizing a shared main expert and selected deputy experts. It exhibits adaptability by tailoring deputy experts to cater to out-of-domain few-shot and zero-shot scenarios. Experimental results on 11 datasets show the superiority of our model compared with recent baselines and LLMs. We also provide statistical and visual evidence of the distinct separation of the two abilities in MoeSumm https://github.com/iriscxy/MoE_Summ

Open-domain question answering (ODQA) has emerged as a pivotal research spotlight in information systems. Existing methods follow two main paradigms to collect evidence: (1) Theretrieve-then-read paradigm retrieves pertinent documents from an external corpus; and (2) thegenerate-then-read paradigm employs large language models (LLMs) to generate relevant documents. However, neither can fully address multifaceted requirements for evidence. To this end, we propose LLMQA, a generalized framework that formulates the ODQA process into three basic steps: query expansion, document selection, and answer generation, combining the superiority of both retrieval-based and generation-based evidence. Since LLMs exhibit their excellent capabilities to accomplish various tasks, we instruct LLMs to play multiple roles as generators, rerankers, and evaluators within our framework, integrating them to collaborate in the ODQA process. Furthermore, we introduce a novel prompt optimization algorithm to refine role-playing prompts and steer LLMs to produce higher-quality evidence and answers. Extensive experimental results on widely used benchmarks (NQ, WebQ, and TriviaQA) demonstrate that LLMQA achieves the best performance in terms of both answer accuracy and evidence quality, showcasing its potential for advancing ODQA research and applications.

Due to the growing importance of using graph neural networks in high-stakes applications, there is a pressing need to interpret the predicted results of these models. Existing methods for explanation have mainly focused on generating sub-graphs comprising important edges for a specific prediction. However, these methods face two issues. Firstly, they lack counterfactual validity as removing the subgraph may not affect the prediction, and generating plausible counterfactual examples has not been adequately explored. Secondly, they cannot be extended to heterogeneous graphs as the complex information involved in such graphs increases the difficulty of generating interpretations. This paper proposes a novel counterfactual learning method, named CF-HGExplainer, for heterogeneous graphs. The method incorporates a semantic-aware attentive pooling strategy for the heterogeneous graph classifier and designs a heterogeneous decision boundaries extraction module to find the common logic for similar graphs based on the extracted graph embeddings from the classifier. Additionally, we propose to greedily perturb nodes and edges based on the distribution of node features and edge plausibility to train a neural network for heterogeneous edge weight learning. Extensive experiments on two public academic datasets demonstrate the effectiveness of CF-HGExplainer compared to state-of-the-art methods on the graph classification task and graph interpretation task.

Automatic summarization plays an important role in the exponential document growth on the Web. On content websites such as CNN.com and WikiHow.com, there often exist various kinds of side information along with the main document for attention attraction and easier understanding, such as videos, images, and queries. Such information can be used for better summarization, as they often explicitly or implicitly mention the essence of the article. However, most of the existing side-aware summarization methods are designed to incorporate either single-modal or multi-modal side information, and cannot effectively adapt to each other. In this paper, we propose a general summarization framework, which can flexibly incorporate various modalities of side information. The main challenges in designing a flexible summarization model with side information include: (1) the side information can be in textual or visualformat, and the model needs to align and unify it with the document into the same semantic space, (2) the side inputs can contain information from variousaspects, and the model should recognize the aspects useful for summarization. To address these two challenges, we first propose a unified topic encoder, which jointly discovers latent topics from the document and various kinds of side information. The learned topics flexibly bridge and guide the information flow between multiple inputs in a graph encoder through a topic-aware interaction. We secondly propose a triplet contrastive learning mechanism to align the single-modal or multi-modal information into a unified semantic space, where thesummary quality is enhanced by better understanding thedocument andside information. Results show that our model significantly surpasses strong baselines on three public single-modal or multi-modal benchmark summarization datasets.

Today, timestamped web documents related to a general news query flood the Internet, and timeline summarization targets this concisely by summarizing the evolution trajectory of events along the timeline. Unlike traditional document summarization, timeline summarization needs to model the time series information of the input events and summarize important events in chronological order. To tackle this challenge, in this article we propose our Unified Timeline Summarizer, which can generate abstractive and extractive timeline summaries in time order. Concretely, in the encoder part, we propose a graph-based event encoder that relates multiple events according to their content dependency and learns a global representation of each event. In the decoder part, to ensure the chronological order of the abstractive summary, we propose to extract the feature of event-level attention in its generation process with sequential information retained and use it to simulate the evolutionary attention of the ground truth summary. The event-level attention can also be used to assist in extracting a summary, where the extracted summary also comes in time sequence. We augment the previous Chinese large-scale timeline summarization dataset and collect a new English timeline dataset. Extensive experiments conducted on these datasets and on the out-of-domain Timeline 17 dataset show that our Unified Timeline Summarizer achieves state-of-the-art performance in terms of both automatic and human evaluations. 1

Machine Learning-as-a-Service systems (MLaaS) have been largely developed for cybersecurity-critical applications, such as detecting network intrusions and fake news campaigns. Despite effectiveness, their robustness against adversarial attacks is one of the key trust concerns for MLaaS deployment. We are thus motivated to assess the adversarial robustness of the Machine Learning models residing at the core of these securitycritical applications with categorical inputs. Previous research efforts on accessing model robustness against manipulation of categorical inputs are specific to use cases and heavily depend on domain knowledge, or require white-box access to the target ML model. Such limitations prevent the robustness assessment from being as a domain-agnostic service provided to various real-world applications. We propose a provably optimal yet computationally highly efficient adversarial robustness assessment protocol for a wide band of ML-driven cybersecurity-critical applications. We demonstrate the use of the domain-agnostic robustness assessment method with substantial experimental study on fake news detection and intrusion detection problems.

Molecular graph generation via deep generative models has attracted increasing attention. This is a challenging problem because it requires optimizing a given objective under a huge search space while obeying the chemical valence rules. Although recently developed molecular generation models have achieved promising results on generating novel, valid and unique molecules, few efforts have been made toward interpretable molecular graph generation. In this work, we propose DEMO, a flow-based model for DisEntangled Molecular graph generatiOn in a completely unsupervised manner, which is able to generate molecular graphs w.r.t. the learned disentangled latent factors that are relevant to molecular semantic features and interpretable structural patterns. Specifically, DEMO is composed of a VAE-encoder and a flow-generator. The VAE-encoder focuses on extracting global features of molecular graphs, and the flow-generator aims at disentangling these features to be corresponding to certain types of understandable molecular structure features while learning data distributions. To generate molecular graphs, DEMO simply runs the flow-generator in the reverse order due to the reversibility of the flow-based models. Extensive experimental results on two benchmark datasets demonstrate that DEMO outperforms the state-of-the-art methods in molecular generation, and takes the first step in interpretable molecular graph generation.

The related work section is an important component of a scientific paper, which highlights the contribution of the target paper in the context of the reference papers. Authors can save their time and effort by using the automatically generated related work section as a draft to complete the final related work. Most of the existing related work section generation methods rely on extracting off-the-shelf sentences to make a comparative discussion about the target work and the reference papers. However, such sentences need to be written in advance and are hard to obtain in practice. Hence, in this paper, we propose an abstractive target-aware related work generator (TAG), which can generate related work sections consisting of new sentences. Concretely, we first propose a target-aware graph encoder, which models the relationships between reference papers and the target paper with target-centered attention mechanisms. In the decoding process, we propose a hierarchical decoder that attends to the nodes of different levels in the graph with keyphrases as semantic indicators. Finally, to generate a more informative related work, we propose multi-level contrastive optimization objectives, which aim to maximize the mutual information between the generated related work with the references and minimize that with non-references. Extensive experiments on two public scholar datasets show that the proposed model brings substantial improvements over several strong baselines in terms of automatic and tailored human evaluations.

Discovering disease-gene association is a fundamental and critical biomedical task, which assists biologists and physicians to discover pathogenic mechanism of syndromes. With various clinical biomarkers measuring the similarities among genes and disease phenotypes, network-based semi-supervised learning (NSSL) has been commonly utilized by these studies to address this class-imbalanced large-scale data issue. However, most existing NSSL approaches are based on linear models and suffer from two major limitations: 1) They implicitly consider a local-structure representation for each candidate; 2) They are unable to capture nonlinear associations between diseases and genes. In this paper, we propose a new framework for disease-gene association task by combining Graph Convolutional Network (GCN) and matrix factorization, named GCN-MF. With the help of GCN, we could capture non-linear interactions and exploit measured similarities. Moreover, we define a margin control loss function to reduce the effect of sparsity. Empirical results demonstrate that the proposed deep learning algorithm outperforms all other state-of-the-art methods on most of metrics.

This paper aims at improvement on the effectiveness of matrix decomposition (MD) methods for implicit feedback. We highlight two critical limitations of existing works. First, due to the large number of unlabeled feedback, most existing works employ a uniform weight to the missing data to reduce computational complexity. However, such a uniform assumption may rarely hold in real-world scenarios. Second, the commonly-used bilateral loss function might be infinite if the data point is mis-classified. Outliers may have such issues and misguide the learning process. We address the above two issues by learning a robust asymmetric learning model. By leveraging the cost-sensitive learning and capped unilateral loss function, our robust MD objective function integrates them into a joint formulation, where the low-rank basis for user/item profiles can be modeled in an effective and robust way. Particularly, a novel log-determinant function is employed to refine the nuclear norm with respect to the low-rank approximation. We derive an iterative re-weighted algorithm to efficiently minimize this MD objective, and also rigorously prove a lower error bound of the proposed algorithm compared to the 1-bit matrix completion method. Finally, we show the promising experimental results of our algorithm on benchmark recommendation datasets.

Recommender systems with implicit feedback (e.g. clicks and purchases) suffer from two critical limitations: 1) imbalanced labels may mislead the learning process of the conventional models that assign balanced weights to the classes; and 2) outliers with large reconstruction errors may dominate the objective function by the conventional $L_2$-norm loss. To address these issues, we propose a robust asymmetric recommendation model. It integrates cost-sensitive learning with capped unilateral loss into a joint objective function, which can be optimized by an iteratively weighted approach. To reduce the computational cost of low-rank approximation, we exploit the dual characterization of the nuclear norm to derive a min-max optimization problem and design a subgradient algorithm without performing full SVD. Finally, promising empirical results demonstrate the effectiveness of our algorithm on benchmark recommendation datasets.

Multi-Task Learning (MTL) can enhance a classifier's generalization performance by learning multiple related tasks simultaneously. Conventional MTL works under the offline or batch setting, and suffers from expensive training cost and poor scalability. To address such inefficiency issues, online learning techniques have been applied to solve MTL problems. However, most existing algorithms of online MTL constrain task relatedness into a presumed structure via a single weight matrix, which is a strict restriction that does not always hold in practice. In this paper, we propose a robust online MTL framework that overcomes this restriction by decomposing the weight matrix into two components: The first one captures the low-rank common structure among tasks via a nuclear norm and the second one identifies the personalized patterns of outlier tasks via a group lasso. Theoretical analysis shows the proposed algorithm can achieve a sub-linear regret with respect to the best linear model in hindsight. Even though the above framework achieves good performance, the nuclear norm that simply adds all nonzero singular values together may not be a good low-rank approximation. To improve the results, we use a log-determinant function as a non-convex rank approximation. The gradient scheme is applied to optimize log-determinant function and can obtain a closed-form solution for this refined problem. Experimental results on a number of real-world applications verify the efficacy of our method.

Multivariate performance measure optimization refers to learning predictive models such that a desired complex performance measure can be optimized over a training set, such as the F1 score. Up to now, all the existing multivariate performance measure optimization methods are limited to a completely labeled data tuple, i.e., the label tuple is complete. However, in real-world applications, sometimes it is difficult to obtain a complete label tuple. In this paper, we show that the multivariate performance measures can also be optimized by learning from partially labeled data tuple, when the label tuple is incomplete. We introduce a slack label tuple to represent the sought complete true label tuple, and learn it jointly with a hyper predictor, so that it can be consistent to the known labels, prediction results, and is smooth in the neighborhood. We develop an iterative learning algorithm to learn the slack label tuple and the hyper predictor. Its advantage over state-of-the-art multivariate performance measure optimization methods is shown by experiments on benchmark data sets.

Many classification algorithms have been successfully deployed in security-sensitive applications including spam filters and intrusion detection systems. Under such adversarial environments, adversaries can generate exploratory attacks against the defender such as evasion and reverse engineering. In this paper, we discuss why reverse engineering attacks can be carried out quite efficiently against fixed classifiers, and investigate the use of randomization as a suitable strategy for mitigating their risk. In particular, we derive a semidefinite programming (SDP) formulation for learning a distribution of classifiers subject to the constraint that any single classifier picked at random from such distribution provides reliable predictions with a high probability. We analyze the tradeoff between variance of the distribution and its predictive accuracy, and establish that one can almost always incorporate randomization with large variance without incurring a loss in accuracy. In other words, the conventional approach of using a fixed classifier in adversarial environments is generally Pareto suboptimal. Finally, we validate such conclusions on both synthetic and real-world classification problems.