The use of Knowledge Graphs (KGs) which constitute large networks of real-world entities and their interrelationships, has grown rapidly. A substantial body of research has emerged, exploring the integration of deep learning (DL) and large language models (LLMs) with KGs. This workshop aims to bring together leading researchers in the field to discuss and foster collaborations on the intersection of KG and DL/LLMs.

Knowledge Graphs (KGs) are a collection of facts describing entities connected by relationships. KG embeddings map entities and relations into a vector space while preserving their relational semantics. This enables effective inference of missing knowledge from the embedding space. Most KG embedding approaches focused on triple-shaped KGs. A great amount of real-world knowledge, however, cannot simply be represented by triples. In this tutorial, we give a systematic introduction to KG embeddings that go beyond the triple representation. In particular, the tutorial will focus on temporal facts where the triples are enriched with temporal information, hyper-relational facts where the triples are enriched with qualifiers, n-ary facts describing relationships between multiple entities, and also facts that are augmented with literal and text descriptions. During the tutorial, we will introduce both fundamental knowledge and advanced topics for understanding recent embedding approaches for beyond-triple representations.

Knowledge Graphs (KG) are of vital importance for multiple applications on the web, including information retrieval, recommender systems, and metadata annotation.

Exponential growths of social media and micro-blogging sites not only provide platforms for empowering freedom of expressions and individual voices, but also enables people to express anti-social behavior like online harassment, cyberbul-lying, and hate speech. Numerous works have been proposed to utilize these data for social and anti-social behavior analysis, document characterization, and sentiment analysis by predicting the contexts mostly for highly resourced languages like English. However, some languages are under-resources, e.g., South Asian languages like Bengali, Tamil, Assamese, Malayalam, that lack of computational resources for natural language processing. In this paper1, we provide several classification benchmarks for Bengali, an under-resourced language. We prepared three datasets of expressing hate, commonly used topics, and opinions for hate speech detection, document classification, and sentiment analysis. We built the largest Bengali word embedding models to date based on 250 million articles, which we call BengFastText. We perform three experiments, covering document classification, sentiment analysis, and hate speech detection. We incorporate word embeddings into a Multichannel Convolutional-LSTM (MC-LSTM) network for predicting different types of hate speech, document classification, and sentiment analysis. Experiments demonstrate that BengFastText can capture the semantics of words from respective contexts correctly. Evaluations against several baseline embedding models, e.g., Word2Vec and GloVe yield up to 92.30%, 82.25%, and 90.45% F1-scores in case of document classification, sentiment analysis, and hate speech detection, respectively during 5-fold cross-validation tests.

Question answering over knowledge graphs (KGQA) has evolved from simple single-fact questions to complex questions that require graph traversal and aggregation. We propose a novel approach for complex KGQA that uses unsupervised message passing, which propagates confidence scores obtained by parsing an input question and matching terms in the knowledge graph to a set of possible answers. First, we identify entity, relationship, and class names mentioned in a natural language question, and map these to their counterparts in the graph. Then, the confidence scores of these mappings propagate through the graph structure to locate the answer entities. Finally, these are aggregated depending on the identified question type. This approach can be efficiently implemented as a series of sparse matrix multiplications mimicking joins over small local subgraphs. Our evaluation results show that the proposed approach outperforms the state of the art on the LC-QuAD benchmark. Moreover, we show that the performance of the approach depends only on the quality of the question interpretation results, i.e., given a correct relevance score distribution, our approach always produces a correct answer ranking. Our error analysis reveals correct answers missing from the benchmark dataset and inconsistencies in the DBpedia knowledge graph. Finally, we provide a comprehensive evaluation of the proposed approach accompanied with an ablation study and an error analysis, which showcase the pitfalls for each of the question answering components in more detail.

Many commonly used data-mining techniques utilized across research fields perform poorly when used for large data sets. Sequential agglomerative hierarchical non-overlapping clustering is one technique for which the algorithms' scaling properties prohibit clustering of a large amount of items. Besides the unfavorable time complexity of O(n2), these algorithms have a space complexity of O(n2), which can be reduced to O(n) if the time complexity is allowed to rise to O(n2 log2n). In this paper, we propose the use of locality-sensitive hashing combined with a novel data structure called twister tries to provide an approximate clustering for average linkage. Our approach requires only linear space. Furthermore, its time complexity is linear in the number of items to be clustered, making it feasible to apply it on a larger scale. We evaluate the approach both analytically and by applying it to several data sets.