QUILL: Quotation Generation Enhancement of Large Language Models

While Large language models (LLMs) have become excellent writing assistants, they still struggle with quotation generation. This is because they either hallucinate when providing factual quotations or fail to provide quotes that exceed human expectations. To bridge the gap, we systematically study how to evaluate and improve LLMs' performance in quotation generation tasks. We first establish a holistic and automatic evaluation system for quotation generation task, which consists of five criteria each with corresponding automatic metric. To improve the LLMs' quotation generation abilities, we construct a bilingual knowledge base that is broad in scope and rich in dimensions, containing up to 32,022 quotes. Moreover, guided by our critiria, we further design a quotation-specific metric to rerank the retrieved quotations from the knowledge base. Extensive experiments show that our metrics strongly correlate with human preferences. Existing LLMs struggle to generate desired quotes, but our quotation knowledge base and reranking metric help narrow this gap. Our dataset and code are publicly available at https://github.com/GraceXiaoo/QUILL.

SEGMENT+: Long Text Processing with Short-Context Language Models

There is a growing interest in expanding the input capacity of language models (LMs) across various domains. However, simply increasing the context window does not guarantee robust performance across diverse long-input processing tasks, such as understanding extensive documents and extracting detailed information from lengthy and noisy data. In response, we introduce SEGMENT+, a general framework that enables LMs to handle extended inputs within limited context windows efficiently. SEGMENT+ utilizes structured notes and a filtering module to manage information flow, resulting in a system that is both controllable and interpretable. Our extensive experiments across various model sizes, focusing on long-document question-answering and Needle-in-a-Haystack tasks, demonstrate the effectiveness of SEGMENT+ in improving performance.

Teaching Large Language Models to Express Knowledge Boundary from Their Own Signals

Large language models (LLMs) have achieved great success, but their occasional content fabrication, or hallucination, limits their practical application. Hallucination arises because LLMs struggle to admit ignorance due to inadequate training on knowledge boundaries. We call it a limitation of LLMs that they can not accurately express their knowledge boundary, answering questions they know while admitting ignorance to questions they do not know. In this paper, we aim to teach LLMs to recognize and express their knowledge boundary, so they can reduce hallucinations caused by fabricating when they do not know. We propose CoKE, which first probes LLMs' knowledge boundary via internal confidence given a set of questions, and then leverages the probing results to elicit the expression of the knowledge boundary. Extensive experiments show CoKE helps LLMs express knowledge boundaries, answering known questions while declining unknown ones, significantly improving in-domain and out-of-domain performance.

SLMRec: Empowering Small Language Models for Sequential Recommendation

The sequential Recommendation (SR) task involves predicting the next item a user is likely to interact with, given their past interactions. The SR models examine the sequence of a user's actions to discern more complex behavioral patterns and temporal dynamics. Recent research demonstrates the great impact of LLMs on sequential recommendation systems, either viewing sequential recommendation as language modeling or serving as the backbone for user representation. Although these methods deliver outstanding performance, there is scant evidence of the necessity of a large language model and how large the language model is needed, especially in the sequential recommendation scene. Meanwhile, due to the huge size of LLMs, it is inefficient and impractical to apply a LLM-based model in real-world platforms that often need to process billions of traffic logs daily. In this paper, we explore the influence of LLMs' depth by conducting extensive experiments on large-scale industry datasets. Surprisingly, we discover that most intermediate layers of LLMs are redundant. Motivated by this insight, we empower small language models for SR, namely SLMRec, which adopt a simple yet effective knowledge distillation method. Moreover, SLMRec is orthogonal to other post-training efficiency techniques, such as quantization and pruning, so that they can be leveraged in combination. Comprehensive experimental results illustrate that the proposed SLMRec model attains the best performance using only 13% of the parameters found in LLM-based recommendation models, while simultaneously achieving up to 6.6x and 8.0x speedups in training and inference time costs, respectively.

Semi-Supervised Clustering with Contrastive Learning for Discovering New Intents

Most dialogue systems in real world rely on predefined intents and answers for QA service, so discovering potential intents from large corpus previously is really important for building such dialogue services. Considering that most scenarios have few intents known already and most intents waiting to be discovered, we focus on semi-supervised text clustering and try to make the proposed method benefit from labeled samples for better overall clustering performance. In this paper, we propose Deep Contrastive Semi-supervised Clustering (DCSC), which aims to cluster text samples in a semi-supervised way and provide grouped intents to operation staff. To make DCSC fully utilize the limited known intents, we propose a two-stage training procedure for DCSC, in which DCSC will be trained on both labeled samples and unlabeled samples, and achieve better text representation and clustering performance. We conduct experiments on two public datasets to compare our model with several popular methods, and the results show DCSC achieve best performance across all datasets and circumstances, indicating the effect of the improvements in our work.

Twitter Bot Detection Using Bidirectional Long Short-term Memory Neural Networks and Word Embeddings

Twitter is a web application playing dual roles of online social networking and micro-blogging. The popularity and open structure of Twitter have attracted a large number of automated programs, known as bots. Legitimate bots generate a large amount of benign contextual content, i.e., tweets delivering news and updating feeds, while malicious bots spread spam or malicious contents. To assist human users in identifying who they are interacting with, this paper focuses on the classification of human and spambot accounts on Twitter, by employing recurrent neural networks, specifically bidirectional Long Short-term Memory (BiLSTM), to efficiently capture features across tweets. To the best of our knowledge, our work is the first that develops a recurrent neural model with word embeddings to distinguish Twitter bots from human accounts, that requires no prior knowledge or assumption about users' profiles, friendship networks, or historical behavior on the target account. Moreover, our model does not require any handcrafted features. The preliminary simulation results are very encouraging. Experiments on the cresci-2017 dataset show that our approach can achieve competitive performance compared with existing state-of-the-art bot detection systems.

OD-GCN object detection by knowledge graph with GCN

Classical object detection frameworks lack of utilizing objects' surrounding information. In this article, we introduce the graph convolutional networks (GCN) into the object detection, and propose a new framework called OD-GCN (object detection with graph convolutional network). It utilizes the category relationship to improve the detection precision. We set up a knowledge graph to reflect the co-exist relationships among objects. GCN plays the role of post-processing to adjust the output of base object detection models. It is a flexible framework that any pre-trained object detection models can be used as the base model. In the experiments, we try several popular base detection models, OD-GCN always improve mAP by 1-5 pp in COCO dataset. In addition, visualized analysis reveals the benchmark improvement is quite logical in human's opinion.

Dual-FOFE-net Neural Models for Entity Linking with PageRank

This paper presents a simple and computationally efficient approach for entity linking (EL), compared with recurrent neural networks (RNNs) or convolutional neural networks (CNNs), by making use of feedforward neural networks (FFNNs) and the recent dual fixed-size ordinally forgetting encoding (dual-FOFE) method to fully encode the sentence fragment and its left/right contexts into a fixed-size representation. Furthermore, in this work, we propose to incorporate PageRank based distillation in our candidate generation module. Our neural linking models consist of three parts: a PageRank based candidate generation module, a dual-FOFE-net neural ranking model and a simple NIL entity clustering system. Experimental results have shown that our proposed neural linking models achieved higher EL accuracy than state-of-the-art models on the TAC2016 task dataset over the baseline system, without requiring any in-house data or complicated handcrafted features. Moreover, it achieves a competitive accuracy on the TAC2017 task dataset.