arXiv subjects

Abdalghani Abujabal

Publications and source records attributed to Abdalghani Abujabal.

7 recordsLinked to original sources

arXiv2026

Redteaming Leading Arabic LLMs with ASAS

As the adoption of large language models (LLMs) grows in Arabic-speaking regions, ensuring their safety and cultural alignment is increasingly critical. However, Arabic LLM safety remains underexplored, especially in adversarial evaluation settings. We introduce the Arabic Safety Index (ASAS), the first fully human-curated Arabic benchmark for redteaming LLMs. ASAS contains 801 prompts spanning 8 safety categories and 8 attack strategies, with ideal responses in Modern Standard Arabic (MSA). We conduct a redteaming evaluation across seven leading models with Arabic capabilities, including GPT-4o, Claude 3.7 Sonnet, and regional models such as ALLaM and FANAR. Human annotators rate responses using a structured 4-point safety scale, revealing that most models fail to defend against 50% of unsafe prompts. Our findings highlight major safety gaps in high-harm categories such as weapons and illicit substances, with direct and obfuscation-based attacks proving most effective. The results also show that language alignment does not readily transfer across languages, and that automated safety judges (e.g., GPT-4o) perform poorly compared to human annotators. ASAS provides a culturally grounded benchmark and redteaming protocol to drive progress in Arabic LLM safety.

cs.AI↗

arXiv2024

ALLaM: Large Language Models for Arabic and English

We present ALLaM: Arabic Large Language Model, a series of large language models to support the ecosystem of Arabic Language Technologies (ALT). ALLaM is carefully trained considering the values of language alignment and knowledge transfer at scale. Our autoregressive decoder-only architecture models demonstrate how second-language acquisition via vocabulary expansion and pretraining on a mixture of Arabic and English text can steer a model towards a new language (Arabic) without any catastrophic forgetting in the original language (English). Furthermore, we highlight the effectiveness of using parallel/translated data to aid the process of knowledge alignment between languages. Finally, we show that extensive alignment with human preferences can significantly enhance the performance of a language model compared to models of a larger scale with lower quality alignment. ALLaM achieves state-of-the-art performance in various Arabic benchmarks, including MMLU Arabic, ACVA, and Arabic Exams. Our aligned models improve both in Arabic and English from their base aligned models.

cs.CL↗

arXiv2021

TEQUILA: Temporal Question Answering over Knowledge Bases

Question answering over knowledge bases (KB-QA) poses challenges in handling complex questions that need to be decomposed into sub-questions. An important case, addressed here, is that of temporal questions, where cues for temporal relations need to be discovered and handled. We present TEQUILA, an enabler method for temporal QA that can run on top of any KB-QA engine. TEQUILA has four stages. It detects if a question has temporal intent. It decomposes and rewrites the question into non-temporal sub-questions and temporal constraints. Answers to sub-questions are then retrieved from the underlying KB-QA engine. Finally, TEQUILA uses constraint reasoning on temporal intervals to compute final answers to the full question. Comparisons against state-of-the-art baselines show the viability of our method.

cs.IR↗

arXiv2020

Answering Complex Questions by Joining Multi-Document Evidence with Quasi Knowledge Graphs

Direct answering of questions that involve multiple entities and relations is a challenge for text-based QA. This problem is most pronounced when answers can be found only by joining evidence from multiple documents. Curated knowledge graphs (KGs) may yield good answers, but are limited by their inherent incompleteness and potential staleness. This paper presents QUEST, a method that can answer complex questions directly from textual sources on-the-fly, by computing similarity joins over partial results from different documents. Our method is completely unsupervised, avoiding training-data bottlenecks and being able to cope with rapidly evolving ad hoc topics and formulation style in user questions. QUEST builds a noisy quasi KG with node and edge weights, consisting of dynamically retrieved entity names and relational phrases. It augments this graph with types and semantic alignments, and computes the best answers by an algorithm for Group Steiner Trees. We evaluate QUEST on benchmarks of complex questions, and show that it substantially outperforms state-of-the-art baselines.

cs.IR↗

arXiv2019

Look before you Hop: Conversational Question Answering over Knowledge Graphs Using Judicious Context Expansion

Fact-centric information needs are rarely one-shot; users typically ask follow-up questions to explore a topic. In such a conversational setting, the user's inputs are often incomplete, with entities or predicates left out, and ungrammatical phrases. This poses a huge challenge to question answering (QA) systems that typically rely on cues in full-fledged interrogative sentences. As a solution, we develop CONVEX: an unsupervised method that can answer incomplete questions over a knowledge graph (KG) by maintaining conversation context using entities and predicates seen so far and automatically inferring missing or ambiguous pieces for follow-up questions. The core of our method is a graph exploration algorithm that judiciously expands a frontier to find candidate answers for the current question. To evaluate CONVEX, we release ConvQuestions, a crowdsourced benchmark with 11,200 distinct conversations from five different domains. We show that CONVEX: (i) adds conversational support to any stand-alone QA system, and (ii) outperforms state-of-the-art baselines and question completion strategies.

cs.IR↗

arXiv2019

Neural Named Entity Recognition from Subword Units

Named entity recognition (NER) is a vital task in spoken language understanding, which aims to identify mentions of named entities in text e.g., from transcribed speech. Existing neural models for NER rely mostly on dedicated word-level representations, which suffer from two main shortcomings. First, the vocabulary size is large, yielding large memory requirements and training time. Second, these models are not able to learn morphological or phonological representations. To remedy the above shortcomings, we adopt a neural solution based on bidirectional LSTMs and conditional random fields, where we rely on subword units, namely characters, phonemes, and bytes. For each word in an utterance, our model learns a representation from each of the subword units. We conducted experiments in a real-world large-scale setting for the use case of a voice-controlled device covering four languages with up to 5.5M utterances per language. Our experiments show that (1) with increasing training data, performance of models trained solely on subword units becomes closer to that of models with dedicated word-level embeddings (91.35 vs 93.92 F1 for English), while using a much smaller vocabulary size (332 vs 74K), (2) subword units enhance models with dedicated word-level embeddings, and (3) combining different subword units improves performance.

cs.CL↗

arXiv2019

ComQA: A Community-sourced Dataset for Complex Factoid Question Answering with Paraphrase Clusters

To bridge the gap between the capabilities of the state-of-the-art in factoid question answering (QA) and what users ask, we need large datasets of real user questions that capture the various question phenomena users are interested in, and the diverse ways in which these questions are formulated. We introduce ComQA, a large dataset of real user questions that exhibit different challenging aspects such as compositionality, temporal reasoning, and comparisons. ComQA questions come from the WikiAnswers community QA platform, which typically contains questions that are not satisfactorily answerable by existing search engine technology. Through a large crowdsourcing effort, we clean the question dataset, group questions into paraphrase clusters, and annotate clusters with their answers. ComQA contains 11,214 questions grouped into 4,834 paraphrase clusters. We detail the process of constructing ComQA, including the measures taken to ensure its high quality while making effective use of crowdsourcing. We also present an extensive analysis of the dataset and the results achieved by state-of-the-art systems on ComQA, demonstrating that our dataset can be a driver of future research on QA.

cs.CL↗