Search

Inspired by the recent advancements of Large Language Models (LLMs) in NLP tasks, there's growing interest in applying LLMs to graph-related tasks. This study delves into the capabilities of instruction-following LLMs for engaging with real-world graphs, aiming to offer empirical insights into how LLMs can effectively interact with graphs and generalize across graph tasks. We begin by constructing a dataset designed for instruction tuning, which comprises a diverse collection of 79 graph-related tasks from academic and e-commerce domains, featuring 44,240 training instances and 18,960 test samples. Utilizing this benchmark, our initial investigation focuses on identifying the optimal graph representation that serves as a conduit for LLMs to understand complex graph structures. Our findings indicate that JSON format for graph representation consistently outperforms natural language and code formats across various LLMs and graph types. Furthermore, we examine the key factors that influence the generalization abilities of instruction-tuned LLMs by evaluating their performance on both in-domain and out-of-domain graph tasks., Comment: COLM 2024

We present a novel system that automatically extracts and generates informative and descriptive sentences from the biomedical corpus and facilitates the efficient search for relational knowledge. Unlike previous search engines or exploration systems that retrieve unconnected passages, our system organizes descriptive sentences as a relational graph, enabling researchers to explore closely related biomedical entities (e.g., diseases treated by a chemical) or indirectly connected entities (e.g., potential drugs for treating a disease). Our system also uses ChatGPT and a fine-tuned relation synthesis model to generate concise and reliable descriptive sentences from retrieved information, reducing the need for extensive human reading effort. With our system, researchers can easily obtain both high-level knowledge and detailed references and interactively steer to the information of interest. We spotlight the application of our system in COVID-19 research, illustrating its utility in areas such as drug repurposing and literature curation., Comment: EMNLP 2023 Demo

In this paper, we propose DimonGen, which aims to generate diverse sentences describing concept relationships in various everyday scenarios. To support this, we first create a benchmark dataset for this task by adapting the existing CommonGen dataset. We then propose a two-stage model called MoREE to generate the target sentences. MoREE consists of a mixture of retrievers model that retrieves diverse context sentences related to the given concepts, and a mixture of generators model that generates diverse sentences based on the retrieved contexts. We conduct experiments on the DimonGen task and show that MoREE outperforms strong baselines in terms of both the quality and diversity of the generated sentences. Our results demonstrate that MoREE is able to generate diverse sentences that reflect different relationships between concepts, leading to a comprehensive understanding of concept relationships., Comment: ACL 2023

We propose DEER (Descriptive Knowledge Graph for Explaining Entity Relationships) - an open and informative form of modeling entity relationships. In DEER, relationships between entities are represented by free-text relation descriptions. For instance, the relationship between entities of machine learning and algorithm can be represented as ``Machine learning explores the study and construction of algorithms that can learn from and make predictions on data.'' To construct DEER, we propose a self-supervised learning method to extract relation descriptions with the analysis of dependency patterns and generate relation descriptions with a transformer-based relation description synthesizing model, where no human labeling is required. Experiments demonstrate that our system can extract and generate high-quality relation descriptions for explaining entity relationships. The results suggest that we can build an open and informative knowledge graph without human annotation., Comment: Accepted to EMNLP 2022

In the third catalog of active galactic nuclei detected by the $Fermi$ Large Area Telescope Clean (3LAC) sample, there are 402 blazars candidates of uncertain type (BCU). The proposed analysis will help to evaluate the potential optical classification flat spectrum radio quasars (FSRQs) versus BL Lacertae (BL Lacs) objects of BCUs, which can help to understand which is the most elusive class of blazar hidden in the Fermi sample. By studying the 3LAC sample, we found some critical values of $\gamma$-ray photon spectral index ($\Gamma_{\rm ph}$), variability index (VI) and radio flux (${\rm F_R}$) of the sources separate known FSRQs and BL Lac objects. We further utilize those values to defined an empirical "high-confidence" candidate zone that can be used to classify the BCUs. Within such a zone ($\Gamma_{\rm ph}<2.187$, log${\rm F_R}<2.258$ and ${ \rm logVI <1.702}$), we found that 120 BCUs can be classified BL Lac candidates with a higher degree of confidence (with a misjudged rate $<1\%$). Our results suggest that an empirical "high confidence" diagnosis is possible to distinguish the BL Lacs from the Fermi observations based on only on the direct observational data of $\Gamma_{\rm ph}$, VI and ${\rm F_R}$., Comment: Published in ApJ, 2020, 891, 87. 11 pages, 1 figures, 5 tables

In the fourth \emph{Fermi} Large Area Telescope source catalog (4FGL), 5064 $\gamma$-ray sources are reported, including 3207 active galactic nuclei (AGNs), 239 pulsars, 1336 unassociated sources, 92 sources with weak association with blazar at low Galactic latitude and 190 other sources. We employ two different supervised machine learning classifiers, combined with the direct observation parameters given by the 4FGL fits table, to search for sources potentially classified as AGNs and pulsars in the 1336 unassociated sources. In order to reduce the error caused by the large difference in the sizes of samples, we divide the classification process into two separate steps in order to identify the AGNs and the pulsars. First, we select the identified AGNs from all of the samples, and then select the identified pulsars from the remaining. Using the 4FGL sources associated or identified as AGNs, pulsars, and other sources with the features selected through the K-S test and the random forest (RF) feature importance measurement, we trained, optimized, and tested our classifier models. Then, the models are applied to classify the 1336 unassociated sources. According to the calculation results of the two classifiers, we show the sensitivity, specificity, accuracy in each step, and the class of unassociated sources given by each classifier. The accuracy obtained in the first step is approximately $95\%$; in the second step, the obtained overall accuracy is approximately $80\%$. Combining the results of the two classifiers, we predict that there are 583 AGN-type candidates, 115 pulsar-type candidates, 154 other types of $\gamma$-ray candidates, and 484 of uncertain types., Comment: Accepted for publication in RAA, 21 pages, 9 figures, 9 tables

The recently published fourth Fermi Large Area Telescope source catalog (4FGL) reports 5065 gamma-ray sources in terms of direct observational gamma-ray properties. Among the sources, the largest population is the Active Galactic Nuclei (AGN), which consists of 3137 blazars, 42 radio galaxies, and 28 other AGNs. The blazar sample comprises 694 flat-spectrum radio quasars (FSRQs), 1131 BL Lac-type objects (BL Lacs), and 1312 blazar candidates of an unknown type (BCUs). The classification of blazars is difficult using optical spectroscopy given the limited knowledge with respect to their intrinsic properties, and the limited availability of astronomical observations. To overcome these challenges, machine learning algorithms are being investigated as alternative approaches. Using the 4FGL catalog, a sample of 3137 Fermi blazars with 23 parameters is systematically selected. Three established supervised machine learning algorithms (random forests (RFs), support vector machines (SVMs), artificial neural networks (ANNs)) are employed to general predictive models to classify the BCUs. We analyze the results for all of the different combinations of parameters. Interestingly, a previously reported trend the use of more parameters leading to higher accuracy is not found. Considering the least number of parameters used, combinations of eight, 12 or 10 parameters in the SVM, ANN, or RF generated models achieve the highest accuracy (Accuracy $\simeq$ 91.8\%, or $\simeq$ 92.9\%). Using the combined classification results from the optimal combinations of parameters, 724 BL Lac type candidates and 332 FSRQ type candidates are predicted; however, 256 remain without a clear prediction., Comment: Accepted for publication in ApJ. 14 pages, 2 figures, 6 tables