--- title: "Shafiq Joty — papers, page 2 · Nanyang Technological University" description: "Shafiq Joty, Nanyang Technological University: 42 papers on Elamur Research (2013–2026), h-index 51 — affiliations, topics, co-authors and…" canonical: "https://research.elamur.ai/free/en/authors/shafiq-joty/page/2/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/authors/shafiq-joty/page/2/) [home](https://md.research.elamur.ai/free/en/) / [authors](https://md.research.elamur.ai/free/en/authors/) / [Shafiq Joty](https://md.research.elamur.ai/free/en/authors/shafiq-joty/) / page 2 # Shafiq Joty · page 2 42 papers in our library · 2013–2026 · page 2 of 2 [The profile, affiliations and co-authors →](https://md.research.elamur.ai/free/en/authors/shafiq-joty/) ## Papers The most cited first (OpenAlex and our library). [← prev](https://md.research.elamur.ai/free/en/authors/shafiq-joty/) [1](https://md.research.elamur.ai/free/en/authors/shafiq-joty/) **2** 26. **[Discourse Structure in Machine Translation Evaluation](https://md.research.elamur.ai/free/en/papers/joty-2017-discourse-structure-machine-translation-evaluation/)** *(2017 · 40 citations)* Standard metrics for machine translation evaluation, such as BLEU, mostly compare surface word overlap. They ignore how clauses within a sentence relate to one another. Joty and colleagues test whether sentence-level discourse structure can close part of this gap. They parse both system outputs and references into… 27. **[Direct Judgement Preference Optimization](https://md.research.elamur.ai/free/en/papers/wang-2025-direct-judgement-preference-optimization/)** *(2025 · 35 citations)* LLM-as-judge models are usually trained with supervised finetuning on fairly small datasets and cover only narrow evaluation formats, often just pairwise comparison. 28. **[Learning Planning-based Reasoning by Trajectories Collection and Process Reward Synthesizing](https://md.research.elamur.ai/free/en/papers/jiao-2024-learning-planning-based-reasoning-trajectories/)** *(2024 · 31 citations)* Planning-style reasoning with LLMs, such as reasoning-as-planning with MCTS, makes rationales more reliable but is slow at inference because every intermediate state must be scored during search. 29. **[L2CEval: Evaluating Language-to-Code Generation Capabilities of Large Language Models](https://md.research.elamur.ai/free/en/papers/ni-2024-l2ceval-evaluating-language-code-generation/)** *(2024 · 27 citations)* Large language models trained on code can turn natural-language requests into executable programs, but earlier evaluations covered isolated tasks, single model families or particular learning setups. 30. **[Speech Act Modeling of Written Asynchronous Conversations with Task-Specific Embeddings and Conditional Structured Models](https://md.research.elamur.ai/free/en/papers/joty-2016-speech-act-modeling-written-asynchronous/)** *(2016 · 27 citations)* The paper addresses sentence-level speech act tagging in written asynchronous conversations such as forum threads and emails. 31. **[Modeling Speech Acts in Asynchronous Conversations: A Neural-CRF Approach](https://md.research.elamur.ai/free/en/papers/joty-2018-modeling-speech-acts-asynchronous-conversations/)** *(2018 · 20 citations)* The article tags sentences of forum and e-mail threads with five speech acts (statement, question, suggestion, polite, response), on two older corpora (TripAdvisor forum, BC3 e-mail) and a new one, QC3, from a community question-answering site (50 threads, two annotators). 32. **[Preference Optimization for Reasoning with Pseudo Feedback](https://md.research.elamur.ai/free/en/papers/jiao-2024-preference-optimization-reasoning-pseudo-feedback/)** *(2024 · 15 citations)* Preference learning methods such as DPO and PPO help LLMs reason better in math and coding. However, they need reliable labels to separate good solutions from bad ones, and expert-verified reasoning data is scarce and costly to scale. The authors propose PFPO (Pseudo-Feedback Preference Optimization), which treats… 33. **[Does Context Matter? ContextualJudgeBench for Evaluating LLM-based Judges in Contextual Settings](https://md.research.elamur.ai/free/en/papers/xu-2025-does-context-matter-contextualjudgebench-evaluating/)** *(2025 · 14 citations)* LLM judges are often marketed as general-purpose evaluators, but they are rarely tested on contextual tasks such as RAG and summarization, where a response must be grounded in supplied external material. 34. **[LLMs Are Biased Towards Output Formats! Systematically Evaluating and Mitigating Output Format Bias of LLMs](https://md.research.elamur.ai/free/en/papers/long-2025-llms-are-biased-output-formats/)** *(2025 · 14 citations)* The paper examines how strongly the quality of LLM answers depends on the output format requested in the prompt. 35. **[Discovering the Gems in Early Layers: Accelerating Long-Context LLMs with 1000x Input Token Reduction](https://md.research.elamur.ai/free/en/papers/shi-2026-discovering-gems-early-layers-accelerating/)** *(2026 · 13 citations)* Long-context LLMs are costly to run. KV-cache eviction methods such as H2O and SnapKV make decoding cheaper, but they leave the prefilling phase untouched, and that phase becomes the main bottleneck as inputs grow. The authors observe that attention in early-to-middle layers already points to the tokens that matter… 36. **[Can We Further Elicit Reasoning in LLMs? Critic-Guided Planning with Retrieval-Augmentation for Solving Challenging Tasks](https://md.research.elamur.ai/free/en/papers/li-2025-can-we-further-elicit-reasoning/)** *(2025 · 13 citations)* LLMs that combine chain-of-thought with retrieval-augmented generation often fail on hard, knowledge-heavy problems such as competitive programming. 37. **[DnA-Eval: Enhancing Large Language Model Evaluation through Decomposition and Aggregation](https://md.research.elamur.ai/free/en/papers/li-2025-dna-eval-enhancing-large-language/)** *(2025 · 10 citations)* Using LLMs as judges of generated text is cheap and scalable, but how reliable they are remains unclear. Standard meta-evaluation asks a model once which of two responses is better and measures agreement with human labels. This hides the reasoning, so a model can match humans by luck. The authors propose DnA-Eval, a… 38. **[Adaptation of Hierarchical Structured Models for Speech Act Recognition in Asynchronous Conversation](https://md.research.elamur.ai/free/en/papers/mohiuddin-2019-adaptation-hierarchical-structured-models-speech/)** *(2019 · 7 citations)* Speech act recognition (SAR) labels each sentence in a conversation with its communicative function, such as Question, Suggestion, Response or Polite. 39. **[J4R: Learning to Judge with Equivalent Initial State Group Relative Policy Optimization](https://md.research.elamur.ai/free/en/papers/xu-2026-j4r-learning-judge-equivalent-initial/)** *(2026 · 3 citations)* LLM-as-judge models handle simple tasks like chat quality well, but they struggle to evaluate reasoning-heavy outputs. 40. **[Hard2Verify: A Step-Level Verification Benchmark for Open-Ended Frontier Math](https://md.research.elamur.ai/free/en/papers/pandit-2026-hard2verify-step-level-verification-benchmark/)** *(2026 · 2 citations)* Frontier LLMs now write Olympiad-level proofs, such as the gold-medal results at IMO 2025. Training and evaluating such systems requires automatic verifiers that can judge each step of a proof, not just the final answer. Existing step-level benchmarks have weaknesses: they use easy questions, outputs from weak… 41. **[Aligning Text, Code, and Vision: A Multi-Objective Reinforcement Learning Framework for Text-to-Visualization](https://md.research.elamur.ai/free/en/papers/rahman-2026-aligning-text-code-vision-multi/)** *(2026)* Text-to-visualization (Text2Vis) means taking a data table and a natural-language question and producing both a short textual answer and executable plotting code. 42. **[Addressing Community Question Answering in English and Arabic](https://md.research.elamur.ai/free/en/papers/martino-2016-addressing-community-question-answering-english/)** *(2016)* The paper tackles question re-ranking in community Question Answering (cQA). When a user posts a new question, the system must find earlier forum questions with the same information need, even though forum text is short and noisy. The authors work on two SemEval-2016 Task 3 subtasks: B (English, Qatar Living forum)… [← prev](https://md.research.elamur.ai/free/en/authors/shafiq-joty/) [1](https://md.research.elamur.ai/free/en/authors/shafiq-joty/) **2**