Besnik Fetahu

Zhiyu Chen, Jason Choi, Besnik Fetahu et al. · amazon-science

Customers interacting with product search engines are increasingly formulating information-seeking queries. Frequently Asked Question (FAQ) retrieval aims to retrieve common question-answer pairs for a user query with question intent. Integrating FAQ retrieval in product search can not only empower users to make more informed purchase decisions, but also enhance user retention through efficient post-purchase support. Determining when an FAQ entry can satisfy a user's information need within product search, without disrupting their shopping experience, represents an important challenge. We propose an intent-aware FAQ retrieval system consisting of (1) an intent classifier that predicts when a user's information need can be answered by an FAQ; (2) a reformulation model that rewrites a query into a natural question. Offline evaluation demonstrates that our approach improves Hit@1 by 13% on retrieving ground-truth FAQs, while reducing latency by 95% compared to baseline systems. These improvements are further validated by real user feedback, where 71% of displayed FAQs on top of product search results received explicit positive user feedback. Overall, our findings show promising directions for integrating FAQ retrieval into product search at scale.

Shervin Malmasi, Anjie Fang, Besnik Fetahu et al.

We present MultiCoNER, a large multilingual dataset for Named Entity Recognition that covers 3 domains (Wiki sentences, questions, and search queries) across 11 languages, as well as multilingual and code-mixing subsets. This dataset is designed to represent contemporary challenges in NER, including low-context scenarios (short and uncased text), syntactically complex entities like movie titles, and long-tail entity distributions. The 26M token dataset is compiled from public resources using techniques such as heuristic-based sentence sampling, template extraction and slotting, and machine translation. We applied two NER models on our dataset: a baseline XLM-RoBERTa model, and a state-of-the-art GEMNET model that leverages gazetteers. The baseline achieves moderate performance (macro-F1=54%), highlighting the difficulty of our data. GEMNET, which uses gazetteers, improvement significantly (average improvement of macro-F1=+30%). MultiCoNER poses challenges even for large pre-trained language models, and we believe that it can help further research in building robust NER systems. MultiCoNER is publicly available at https://registry.opendata.aws/multiconer/ and we hope that this resource will help advance research in various aspects of NER.

Besnik Fetahu, Pedro Faustini, Giuseppe Castellucci et al. · amazon-science

The adoption of voice assistants like Alexa or Siri has grown rapidly, allowing users to instantly access information via voice search. Query suggestion is a standard feature of screen-based search experiences, allowing users to explore additional topics. However, this is not trivial to implement in voice-based settings. To enable this, we tackle the novel task of suggesting questions with compact and natural voice hints to allow users to ask follow-up questions. We define the task, ground it in syntactic theory and outline linguistic desiderata for spoken hints. We propose baselines and an approach using sequence-to-sequence Transformers to generate spoken hints from a list of questions. Using a new dataset of 6681 input questions and human written hints, we evaluated the models with automatic metrics and human evaluation. Results show that a naive approach of concatenating suggested questions creates poor voice hints. Our approach, which applies a linguistically-motivated pretraining task was strongly preferred by humans for producing the most natural hints.

Zhiyu Chen, Jie Zhao, Anjie Fang et al.

Conversational Question Answering (CQA) aims to answer questions contained within dialogues, which are not easily interpretable without context. Developing a model to rewrite conversational questions into self-contained ones is an emerging solution in industry settings as it allows using existing single-turn QA systems to avoid training a CQA model from scratch. Previous work trains rewriting models using human rewrites as supervision. However, such objectives are disconnected with QA models and therefore more human-like rewrites do not guarantee better QA performance. In this paper we propose using QA feedback to supervise the rewriting model with reinforcement learning. Experiments show that our approach can effectively improve QA performance over baselines for both extractive and retrieval QA. Furthermore, human evaluation shows that our method can generate more accurate and detailed rewrites when compared to human annotations.

Besnik Fetahu, Zhiyu Chen, Sudipta Kar et al.

We present MULTICONER V2, a dataset for fine-grained Named Entity Recognition covering 33 entity classes across 12 languages, in both monolingual and multilingual settings. This dataset aims to tackle the following practical challenges in NER: (i) effective handling of fine-grained classes that include complex entities like movie titles, and (ii) performance degradation due to noise generated from typing mistakes or OCR errors. The dataset is compiled from open resources like Wikipedia and Wikidata, and is publicly available. Evaluation based on the XLM-RoBERTa baseline highlights the unique challenges posed by MULTICONER V2: (i) the fine-grained taxonomy is challenging, where the scores are low with macro-F1=0.63 (across all languages), and (ii) the corruption strategy significantly impairs performance, with entity corruption resulting in 9% lower performance relative to non-entity corruptions across all languages. This highlights the greater impact of entity noise in contrast to context noise.

Besnik Fetahu, Zhiyu Chen, Oleg Rokhlenko et al.

E-commerce product catalogs contain billions of items. Most products have lengthy titles, as sellers pack them with product attributes to improve retrieval, and highlight key product aspects. This results in a gap between such unnatural products titles, and how customers refer to them. It also limits how e-commerce stores can use these seller-provided titles for recommendation, QA, or review summarization. Inspired by recent work on instruction-tuned LLMs, we present InstructPTS, a controllable approach for the task of Product Title Summarization (PTS). Trained using a novel instruction fine-tuning strategy, our approach is able to summarize product titles according to various criteria (e.g. number of words in a summary, inclusion of specific phrases, etc.). Extensive evaluation on a real-world e-commerce catalog shows that compared to simple fine-tuning of LLMs, our proposed approach can generate more accurate product name summaries, with an improvement of over 14 and 8 BLEU and ROUGE points, respectively.

Chunyuan Deng, Sanket Lokegaonkar, Colin Lockard et al. · gatech

Modern language models still rely on fixed, pre-defined subword tokenizations. Once a tokenizer is trained, the LM can only operate at this fixed level of granularity, which often leads to brittle and counterintuitive behaviors even in otherwise strong reasoning models. We introduce \textbf{ByteFlow Net}, a new hierarchical architecture that removes tokenizers entirely and instead enables models to learn their own segmentation of raw byte streams into semantically meaningful units. ByteFlow Net performs compression-driven segmentation based on the coding rate of latent representations, yielding adaptive boundaries \emph{while preserving a static computation graph via Top-$K$ selection}. Unlike prior self-tokenizing methods that depend on brittle heuristics with human-designed inductive biases, ByteFlow Net adapts its internal representation granularity to the input itself. Experiments demonstrate that this compression-based chunking strategy yields substantial performance gains, with ByteFlow Net outperforming both BPE-based Transformers and previous byte-level architectures. These results suggest that end-to-end, tokenizer-free modeling is not only feasible but also more effective, opening a path toward more adaptive and information-grounded language models.

Xiangci Li, Zhiyu Chen, Jason Ingyu Choi et al.

The goal of conversational product search (CPS) is to develop an intelligent, chat-based shopping assistant that can directly interact with customers to understand shopping intents, ask clarification questions, and find relevant products. However, training such assistants is hindered mainly due to the lack of reliable and large-scale datasets. Prior human-annotated CPS datasets are extremely small in size and lack integration with real-world product search systems. We propose a novel approach, TRACER, which leverages large language models (LLMs) to generate realistic and natural conversations for different shopping domains. TRACER's novelty lies in grounding the generation to dialogue plans, which are product search trajectories predicted from a decision tree model, that guarantees relevant product discovery in the shortest number of search conditions. We also release the first target-oriented CPS dataset Wizard of Shopping (WoS), containing highly natural and coherent conversations (3.6k) from three shopping domains. Finally, we demonstrate the quality and effectiveness of WoS via human evaluations and downstream tasks.

Zhiyu Chen, Jason Choi, Besnik Fetahu et al.

In e-commerce, high consideration search missions typically require careful and elaborate decision making, and involve a substantial research investment from customers. We consider the task of identifying High Consideration (HC) queries. Identifying such queries enables e-commerce sites to better serve user needs using targeted experiences such as curated QA widgets that help users reach purchase decisions. We explore the task by proposing an Engagement-based Query Ranking (EQR) approach, focusing on query ranking to indicate potential engagement levels with query-related shopping knowledge content during product search. Unlike previous studies on predicting trends, EQR prioritizes query-level features related to customer behavior, finance, and catalog information rather than popularity signals. We introduce an accurate and scalable method for EQR and present experimental results demonstrating its effectiveness. Offline experiments show strong ranking performance. Human evaluation shows a precision of 96% for HC queries identified by our model. The model was commercially deployed, and shown to outperform human-selected queries in terms of downstream customer impact, as measured through engagement.

Lütfi Kerem Senel, Besnik Fetahu, Davis Yoshida et al.

Recommender systems are widely used to suggest engaging content, and Large Language Models (LLMs) have given rise to generative recommenders. Such systems can directly generate items, including for open-set tasks like question suggestion. While the world knowledge of LLMs enable good recommendations, improving the generated content through user feedback is challenging as continuously fine-tuning LLMs is prohibitively expensive. We present a training-free approach for optimizing generative recommenders by connecting user feedback loops to LLM-based optimizers. We propose a generative explore-exploit method that can not only exploit generated items with known high engagement, but also actively explore and discover hidden population preferences to improve recommendation quality. We evaluate our approach on question generation in two domains (e-commerce and general knowledge), and model user feedback with Click Through Rate (CTR). Experiments show our LLM-based explore-exploit approach can iteratively improve recommendations, and consistently increase CTR. Ablation analysis shows that generative exploration is key to learning user preferences, avoiding the pitfalls of greedy exploit-only approaches. A human evaluation strongly supports our quantitative findings.

Besnik Fetahu, Nachshon Cohen, Elad Haramaty et al.

Voice assistants have become ubiquitous in smart devices allowing users to instantly access information via voice questions. While extensive research has been conducted in question answering for voice search, little attention has been paid on how to enable proactive recommendations from a voice assistant to its users. This is a highly challenging problem that often leads to user friction, mainly due to recommendations provided to the users at the wrong time. We focus on the domain of e-commerce, namely in identifying Shopping Product Questions (SPQs), where the user asking a product-related question may have an underlying shopping need. Identifying a user's shopping need allows voice assistants to enhance shopping experience by determining when to provide recommendations, such as product or deal recommendations, or proactive shopping actions recommendation. Identifying SPQs is a challenging problem and cannot be done from question text alone, and thus requires to infer latent user behavior patterns inferred from user's past shopping history. We propose features that capture the user's latent shopping behavior from their purchase history, and combine them using a novel Mixture-of-Experts (MoE) model. Our evaluation shows that the proposed approach is able to identify SPQs with a high score of F1=0.91. Furthermore, based on an online evaluation with real voice assistant users, we identify SPQs in real-time and recommend shopping actions to users to add the queried product into their shopping list. We demonstrate that we are able to accurately identify SPQs, as indicated by the significantly higher rate of added products to users' shopping lists when being prompted after SPQs vs random PQs.

Besnik Fetahu, Tejas Mehta, Qun Song et al.

E-commerce customers frequently seek detailed product information for purchase decisions, commonly contacting sellers directly with extended queries. This manual response requirement imposes additional costs and disrupts buyer's shopping experience with response time fluctuations ranging from hours to days. We seek to automate buyer inquiries to sellers in a leading e-commerce store using a domain-specific federated Question Answering (QA) system. The main challenge is adapting current QA systems, designed for single questions, to address detailed customer queries. We address this with a low-latency, sequence-to-sequence approach, MESSAGE-TO-QUESTION ( M2Q ). It reformulates buyer messages into succinct questions by identifying and extracting the most salient information from a message. Evaluation against baselines shows that M2Q yields relative increases of 757% in question understanding, and 1,746% in answering rate from the federated QA system. Live deployment shows that automatic answering saves sellers from manually responding to millions of messages per year, and also accelerates customer purchase decisions by eliminating the need for buyers to wait for a reply

Lingbo Mo, Besnik Fetahu, Oleg Rokhlenko et al.

Yes/No or polar questions represent one of the main linguistic question categories. They consist of a main interrogative clause, for which the answer is binary (assertion or negation). Polar questions and answers (PQA) represent a valuable knowledge resource present in many community and other curated QA sources, such as forums or e-commerce applications. Using answers to polar questions alone in other contexts is not trivial. Answers are contextualized, and presume that the interrogative question clause and any shared knowledge between the asker and answerer are provided. We address the problem of controllable rewriting of answers to polar questions into decontextualized and succinct factual statements. We propose a Transformer sequence to sequence model that utilizes soft-constraints to ensure controllable rewriting, such that the output statement is semantically equivalent to its PQA input. Evaluation on three separate PQA datasets as measured through automated and human evaluation metrics show that our proposed approach achieves the best performance when compared to existing baselines.

Pedro Faustini, Zhiyu Chen, Besnik Fetahu et al.

Spoken Question Answering (QA) is a key feature of voice assistants, usually backed by multiple QA systems. Users ask questions via spontaneous speech which can contain disfluencies, errors, and informal syntax or phrasing. This is a major challenge in QA, causing unanswered questions or irrelevant answers, and leading to bad user experiences. We analyze failed QA requests to identify core challenges: lexical gaps, proposition types, complex syntactic structure, and high specificity. We propose a Semantic Question Reformulation (SURF) model offering three linguistically-grounded operations (repair, syntactic reshaping, generalization) to rewrite questions to facilitate answering. Offline evaluation on 1M unanswered questions from a leading voice assistant shows that SURF significantly improves answer rates: up to 24% of previously unanswered questions obtain relevant answers (75%). Live deployment shows positive impact for millions of customers with unanswered questions; explicit relevance feedback shows high user satisfaction.

Besnik Fetahu, Sudipta Kar, Zhiyu Chen et al.

We present the findings of SemEval-2023 Task 2 on Fine-grained Multilingual Named Entity Recognition (MultiCoNER 2). Divided into 13 tracks, the task focused on methods to identify complex fine-grained named entities (like WRITTENWORK, VEHICLE, MUSICALGRP) across 12 languages, in both monolingual and multilingual scenarios, as well as noisy settings. The task used the MultiCoNER V2 dataset, composed of 2.2 million instances in Bangla, Chinese, English, Farsi, French, German, Hindi, Italian., Portuguese, Spanish, Swedish, and Ukrainian. MultiCoNER 2 was one of the most popular tasks of SemEval-2023. It attracted 842 submissions from 47 teams, and 34 teams submitted system papers. Results showed that complex entity types such as media titles and product names were the most challenging. Methods fusing external knowledge into transformer models achieved the best performance, and the largest gains were on the Creative Work and Group classes, which are still challenging even with external knowledge. Some fine-grained classes proved to be more challenging than others, such as SCIENTIST, ARTWORK, and PRIVATECORP. We also observed that noisy data has a significant impact on model performance, with an average drop of 10% on the noisy subset. The task highlights the need for future research on improving NER robustness on noisy data containing complex entities.

Nachshon Cohen, Amit Portnoy, Besnik Fetahu et al.

BERT based ranking models have achieved superior performance on various information retrieval tasks. However, the large number of parameters and complex self-attention operation come at a significant latency overhead. To remedy this, recent works propose late-interaction architectures, which allow pre-computation of intermediate document representations, thus reducing the runtime latency. Nonetheless, having solved the immediate latency issue, these methods now introduce storage costs and network fetching latency, which limits their adoption in real-life production systems. In this work, we propose the Succinct Document Representation (SDR) scheme that computes highly compressed intermediate document representations, mitigating the storage/network issue. Our approach first reduces the dimension of token representations by encoding them using a novel autoencoder architecture that uses the document's textual content in both the encoding and decoding phases. After this token encoding step, we further reduce the size of entire document representations using a modern quantization technique. Extensive evaluations on passage re-reranking on the MSMARCO dataset show that compared to existing approaches using compressed document representations, our method is highly efficient, achieving 4x-11.6x better compression rates for the same ranking quality.

Lijun Lyu, Maria Koutraki, Martin Krickl et al.

Optical character recognition (OCR) is crucial for a deeper access to historical collections. OCR needs to account for orthographic variations, typefaces, or language evolution (i.e., new letters, word spellings), as the main source of character, word, or word segmentation transcription errors. For digital corpora of historical prints, the errors are further exacerbated due to low scan quality and lack of language standardization. For the task of OCR post-hoc correction, we propose a neural approach based on a combination of recurrent (RNN) and deep convolutional network (ConvNet) to correct OCR transcription errors. At character level we flexibly capture errors, and decode the corrected output based on a novel attention mechanism. Accounting for the input and output similarity, we propose a new loss function that rewards the model's correcting behavior. Evaluation on a historical book corpus in German language shows that our models are robust in capturing diverse OCR transcription errors and reduce the word error rate of 32.3% by more than 89%.

Vasileios Iosifidis, Besnik Fetahu, Eirini Ntoutsi

Automated decision making based on big data and machine learning (ML) algorithms can result in discriminatory decisions against certain protected groups defined upon personal data like gender, race, sexual orientation etc. Such algorithms designed to discover patterns in big data might not only pick up any encoded societal biases in the training data, but even worse, they might reinforce such biases resulting in more severe discrimination. The majority of thus far proposed fairness-aware machine learning approaches focus solely on the pre-, in- or post-processing steps of the machine learning process, that is, input data, learning algorithms or derived models, respectively. However, the fairness problem cannot be isolated to a single step of the ML process. Rather, discrimination is often a result of complex interactions between big data and algorithms, and therefore, a more holistic approach is required. The proposed FAE (Fairness-Aware Ensemble) framework combines fairness-related interventions at both pre- and postprocessing steps of the data analysis process. In the preprocessing step, we tackle the problems of under-representation of the protected group (group imbalance) and of class-imbalance by generating balanced training samples. In the post-processing step, we tackle the problem of class overlapping by shifting the decision boundary in the direction of fairness.

Miriam Redi, Besnik Fetahu, Jonathan Morgan et al.

Wikipedia is playing an increasingly central role on the web,and the policies its contributors follow when sourcing and fact-checking content affect million of readers. Among these core guiding principles, verifiability policies have a particularly important role. Verifiability requires that information included in a Wikipedia article be corroborated against reliable secondary sources. Because of the manual labor needed to curate and fact-check Wikipedia at scale, however, its contents do not always evenly comply with these policies. Citations (i.e. reference to external sources) may not conform to verifiability requirements or may be missing altogether, potentially weakening the reliability of specific topic areas of the free encyclopedia. In this paper, we aim to provide an empirical characterization of the reasons why and how Wikipedia cites external sources to comply with its own verifiability guidelines. First, we construct a taxonomy of reasons why inline citations are required by collecting labeled data from editors of multiple Wikipedia language editions. We then collect a large-scale crowdsourced dataset of Wikipedia sentences annotated with categories derived from this taxonomy. Finally, we design and evaluate algorithmic models to determine if a statement requires a citation, and to predict the citation reason based on our taxonomy. We evaluate the robustness of such models across different classes of Wikipedia articles of varying quality, as well as on an additional dataset of claims annotated for fact-checking purposes.

Christoph Hube, Besnik Fetahu

Biased language commonly occurs around topics which are of controversial nature, thus, stirring disagreement between the different involved parties of a discussion. This is due to the fact that for language and its use, specifically, the understanding and use of phrases, the stances are cohesive within the particular groups. However, such cohesiveness does not hold across groups. In collaborative environments or environments where impartial language is desired (e.g. Wikipedia, news media), statements and the language therein should represent equally the involved parties and be neutrally phrased. Biased language is introduced through the presence of inflammatory words or phrases, or statements that may be incorrect or one-sided, thus violating such consensus. In this work, we focus on the specific case of phrasing bias, which may be introduced through specific inflammatory words or phrases in a statement. For this purpose, we propose an approach that relies on a recurrent neural networks in order to capture the inter-dependencies between words in a phrase that introduced bias. We perform a thorough experimental evaluation, where we show the advantages of a neural based approach over competitors that rely on word lexicons and other hand-crafted features in detecting biased language. We are able to distinguish biased statements with a precision of P=0.92, thus significantly outperforming baseline models with an improvement of over 30%. Finally, we release the largest corpus of statements annotated for biased language.

Besnik Fetahu

Verifiability is one of the core editing principles in Wikipedia, where editors are encouraged to provide citations for the added statements. Statements can be any arbitrary piece of text, ranging from a sentence up to a paragraph. However, in many cases, citations are either outdated, missing, or link to non-existing references (e.g. dead URL, moved content etc.). In total, 20\% of the cases such citations refer to news articles and represent the second most cited source. Even in cases where citations are provided, there are no explicit indicators for the span of a citation for a given piece of text. In addition to issues related with the verifiability principle, many Wikipedia entity pages are incomplete, with relevant information that is already available in online news sources missing. Even for the already existing citations, there is often a delay between the news publication time and the reference time. In this thesis, we address the aforementioned issues and propose automated approaches that enforce the verifiability principle in Wikipedia, and suggest relevant and missing news references for further enriching Wikipedia entity pages.

Besnik Fetahu, Katja Markert, Avishek Anand

\emph{Verifiability} is one of the core editing principles in Wikipedia, editors being encouraged to provide citations for the added content. For a Wikipedia article, determining the \emph{citation span} of a citation, i.e. what content is covered by a citation, is important as it helps decide for which content citations are still missing. We are the first to address the problem of determining the \emph{citation span} in Wikipedia articles. We approach this problem by classifying which textual fragments in an article are covered by a citation. We propose a sequence classification approach where for a paragraph and a citation, we determine the citation span at a fine-grained level. We provide a thorough experimental evaluation and compare our approach against baselines adopted from the scientific domain, where we show improvement for all evaluation metrics.

Besnik Fetahu, Ujwal Gadiraju, Stefan Dietze

The increasing amount of data on the Web, in particular of Linked Data, has led to a diverse landscape of datasets, which make entity retrieval a challenging task. Explicit cross-dataset links, for instance to indicate co-references or related entities can significantly improve entity retrieval. However, only a small fraction of entities are interlinked through explicit statements. In this paper, we propose a two-fold entity retrieval approach. In a first, offline preprocessing step, we cluster entities based on the \emph{x--means} and \emph{spectral} clustering algorithms. In the second step, we propose an optimized retrieval model which takes advantage of our precomputed clusters. For a given set of entities retrieved by the BM25F retrieval approach and a given user query, we further expand the result set with relevant entities by considering features of the queries, entities and the precomputed clusters. Finally, we re-rank the expanded result set with respect to the relevance to the query. We perform a thorough experimental evaluation on the Billions Triple Challenge (BTC12) dataset. The proposed approach shows significant improvements compared to the baseline and state of the art approaches.

Besnik Fetahu, Abhijit Anand, Avishek Anand

Wikipedia, rich in entities and events, is an invaluable resource for various knowledge harvesting, extraction and mining tasks. Numerous resources like DBpedia, YAGO and other knowledge bases are based on extracting entity and event based knowledge from it. Online news, on the other hand, is an authoritative and rich source for emerging entities, events and facts relating to existing entities. In this work, we study the creation of entities in Wikipedia with respect to news by studying how entity and event based information flows from news to Wikipedia. We analyze the lag of Wikipedia (based on the revision history of the English Wikipedia) with 20 years of \emph{The New York Times} dataset (NYT). We model and analyze the lag of entities and events, namely their first appearance in Wikipedia and in NYT, respectively. In our extensive experimental analysis, we find that almost 20\% of the external references in entity pages are news articles encoding the importance of news to Wikipedia. Second, we observe that the entity-based lag follows a normal distribution with a high standard deviation, whereas the lag for news-based events is typically very low. Finally, we find that events are responsible for creation of emergent entities with as many as 12\% of the entities mentioned in the event page are created after the creation of the event page.

Besnik Fetahu, Katja Markert, Avishek Anand

Wikipedia entity pages are a valuable source of information for direct consumption and for knowledge-base construction, update and maintenance. Facts in these entity pages are typically supported by references. Recent studies show that as much as 20\% of the references are from online news sources. However, many entity pages are incomplete even if relevant information is already available in existing news articles. Even for the already present references, there is often a delay between the news article publication time and the reference time. In this work, we therefore look at Wikipedia through the lens of news and propose a novel news-article suggestion task to improve news coverage in Wikipedia, and reduce the lag of newsworthy references. Our work finds direct application, as a precursor, to Wikipedia page generation and knowledge-base acceleration tasks that rely on relevant and high quality input sources. We propose a two-stage supervised approach for suggesting news articles to entity pages for a given state of Wikipedia. First, we suggest news articles to Wikipedia entities (article-entity placement) relying on a rich set of features which take into account the \emph{salience} and \emph{relative authority} of entities, and the \emph{novelty} of news articles to entity pages. Second, we determine the exact section in the entity page for the input article (article-section placement) guided by class-based section templates. We perform an extensive evaluation of our approach based on ground-truth data that is extracted from external references in Wikipedia. We achieve a high precision value of up to 93\% in the \emph{article-entity} suggestion stage and upto 84\% for the \emph{article-section placement}. Finally, we compare our approach against competitive baselines and show significant improvements.

Besnik Fetahu, Katja Markert, Wolfgang Nejdl et al.

An important editing policy in Wikipedia is to provide citations for added statements in Wikipedia pages, where statements can be arbitrary pieces of text, ranging from a sentence to a paragraph. In many cases citations are either outdated or missing altogether. In this work we address the problem of finding and updating news citations for statements in entity pages. We propose a two-stage supervised approach for this problem. In the first step, we construct a classifier to find out whether statements need a news citation or other kinds of citations (web, book, journal, etc.). In the second step, we develop a news citation algorithm for Wikipedia statements, which recommends appropriate citations from a given news collection. Apart from IR techniques that use the statement to query the news collection, we also formalize three properties of an appropriate citation, namely: (i) the citation should entail the Wikipedia statement, (ii) the statement should be central to the citation, and (iii) the citation should be from an authoritative source. We perform an extensive evaluation of both steps, using 20 million articles from a real-world news collection. Our results are quite promising, and show that we can perform this task with high precision and at scale.