United States

Automatic text simplification (ATS) aims to en hance language accessibility for various target groups, particularly persons with intellectual disabilities. Recent advancements in large lan guage models (LLMs) have substantially im proved the quality of machine-generated text simplifications, however, existing LLM-based ATS systems do not incorporate preference feedback during post-training, resulting in a lack of personalization tailored to the specific needs of target group persons. In this work, we propose an ATS personalization framework using direct preference optimization (DPO). Specifically, we post-trained LLM-based ATS models using human feedback collected from persons with intellectual disabilities, reflecting their preferences of paired text simplifications generated by mainstream LLMs. Our pipeline for developing personalized LLM-based ATS systems encompasses data collection, model se lection, supervised fine-tuning (SFT) and DPO post-training, and result evaluation. Our find ings underscore the necessity of active partici pation of target group persons in designing per sonalized inclusive AI solutions aligned with human preferences.

ACL 2026

Evaluating Direct Preference Optimization for Personalizing German Automatic Text Simplifications for Persons with Intellectual Disabilities

language accessibility

personalization

text simplification

## Welcome to EMNLP 2022!
I am delighted to welcome you to EMNLP 2022! I believe this conference has been complicated beyond any precedent. Over the past year, it’s been thrilling to see the organization team approach each new puzzle with creativity and enthusiasm. We hope that those participating in Abu Dhabi as well as those joining remotely will leave the conference feeling newly inspired by the program and newly connected to our ever-growing community. Following EMNLP 2021 and major NLP conferences since, EMNLP 2022 is “hybrid,” serving both virtual and in-person participants.

Our key innovations for EMNLP 2022 include:

* EMNLP 2022 is “hybrid” in a second sense, as well: we allowed both direct and rolling review paper submissions, building on the pilot experiment of EMNLP 2021, which considered a small number of ARR submissions. 
* Familiar from NAACL but new to EMNLP, we’ve added an industry track.
* During the conference, “portals” will link virtual poster sessions to in-person conference participants during poster sessions each day.
* The first ACL-family conference in the United Arab Emirates.

 *Message from Noah A. Smith, University of Washington and Allen Institute for AI, Seattle, Washington, USA* 
***EMNLP 2022 General Chair***
 
[![](https://assets.underline.io/uploads/markdown_image/1/image/9eec7d4a287ee18c278b08229290aa83.png)](https://drive.google.com/file/d/1OlPv6QBeo62VVTughj2jkiLeyHd1WnUt/view)
 
[![](https://assets.underline.io/uploads/markdown_image/1/image/a3db7a768409f05192210d98601edb25.png)](https://emnlp2022.rocket.chat/)

To access this site you need to register. Please register [here](https://2022.emnlp.org/registration/).

Register here

EMNLP 2022

Welcome!
EMNLP 2022 will take place in Abu Dhabi from December 7th to December 11th, 2022. And it will be held in hybrid mode, both online and offline.

workshop paper

It is my great pleasure and honor to welcome you to the 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026), held in San Diego, California, USA, July 2–7, 2026. ACL 2026 continues our fields tradition of excellence in scholarship, innovation, and inclusivity, and I am deeply grateful to the many volunteers who have worked tirelessly to bring this event to life.

It is exciting to see the many developments in the field reflected at our conference, such as the advancement of large language models (born out of our community) into many application directions. This is not without challenges, partly due to unsolved problems, and partly due to misuse of the technology. All this is reflected well at ACL 2026. The conference also reflects that the field is not only about advancing technology, but also the science of fundamentally understanding language and how it is use by many different speaker communities. 
[Full Message from General Chair Philipp Koehn](https://drive.google.com/file/d/1QiJFOmMjvDgsDo3i1XzlQnImxe3RU_Cy/view?usp=sharing) 
<html><button style="display: inline-flex; align-items: center; justify-content: center; white-space: nowrap; border-radius: 9999px; font-weight: bold; background: #7c3aed; color: white; font-family: 'Space Grotesk', sans-serif; height: 40px; font-size: 16px; padding: 0 20px; border: none; cursor: pointer" onclick="window.open('https://drive.google.com/file/d/1QiJFOmMjvDgsDo3i1XzlQnImxe3RU_Cy/view?usp=sharing')">Conference Handbook</button></html>

To access this event page, you need to log in with the **email address you registered with**. Access credentials will be sent to your email from Underline - subject line "Welcome to ACL 2026". Please be sure to check your spam email folder if you do not see an email confirmation right away.

To access this event page, you are required to register.
Please complete your registration to continue.

We recommend reading [**the registration information**](https://2026.aclweb.org/registration/) first.

Registration Required

The 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026) will take place in San Diego, California, from July 2 - 7, 2026.

This paper presents a methodology that uses LLMs to align multilingual offensive lexi cons at the sense level. Lexicons of differ ent structures and origins in Arabic, Bulgar ian, Modern Greek, French, and Italian have been aligned directly without pivoting through English. The Modern Greek lexicon is LLM generated, and the other four lexicons are WordNet-compatible. For inter-language align ment of senses, an LLM-as-a-judge rubric was used over lemma–definition–example triples. The LLM makes 2.87M pairwise comparisons and yields 31 strict global-sense categories. The paper discusses the challenges involved in sense alignment tasks. The resource is avail able to support downstream applications such as Machine Translation and cross-lingual hate speech detection.

From Form to Meaning: Interlingua Sense-Alignment of Offensive Language with LLMs

Large Language Models (LLMs) perform strongly on many language tasks, but their capa bility in structurally constrained, accessibility critical modalities such as Braille remains un clear. We evaluate state-of-the-art LLMs on bidirectional Korean–Braille translation using a human-annotated dataset. Despite expecta tions that multilingual, instruction-tuned mod els can generalize to Braille via text repre sentations, we find consistently poor, unstable outputs and substantial disagreement with hu man judgments. These results point to missing Braille-aware tokenization and weak alignment between Korean and Braille patterns. In con trast, supervised fine-tuning of a small model (T5-small) on the same data yields large and stable gains over zero-shot and prompted LLM baselines across standard metrics (SacreBLEU, ChrF++, CER, BLEU, ROUGE-L, METEOR, CIDEr). Our findings reveal a systematic lim itation of current LLMs and demonstrate the effectiveness of modest task-specific supervi sion.

I'm Sorry, but I Can't Help with Braille: Revealing Accessibility Failures in State-of-the-Art LLMs

We investigate the role of large language mod els (LLMs) in promoting gender-inclusive lan guage by evaluating their ability to rewrite bi ased text and generate counterfactual narratives across multiple languages. We introduce a shared task with two subtasks: gender-inclusive rewriting and counterfactual generation. The task covers five languages English, German, Spanish, Tamil, and Kannada reflecting diverse grammatical gender systems and sociocultural contexts. We release curated word-level and sentence-level datasets to support controlled inclusive generation. A total of 50 teams reg istered for the shared task, and around 8 teams submitted results. Submissions are evaluated using a hybrid framework combining rubric based automatic scoring with expert human judgment. Finally, we provide an overview of participating systems and discuss key findings and challenges observed across languages.

Insights from Multilingual Gender Inclusive Language Generation Shared Task

Gender bias in multilingual language gener ation systems poses serious ethical and so cial issues, especially in languages with com plex morphology. In this study, we propose a lightweight multilingual approach that employs instruction-guided fine-tuning of the mT5 small transformer model for gender-inclusive language generation. The framework accom modates five languages: English, German, Spanish, Tamil, and Kannada. The approach uses the task-prefix rewriting method to trans form gender-specific sentences to their gender neutral versions. The training data from differ ent languages is combined into a single multi lingual dataset for sequence-to-sequence fine tuning. Beam search decoding with repetition constraints is used during inference to improve the quality of the output. The system’s perfor mance is measured using GIFI, semantic simi larity, and an overall combined score across all languages. Experimental results show that the system can eliminate gender-biased language while retaining semantic meaning in part across languages.

CAI@LTEDI 2026: Multilingual Gender Inclusive Language Generation using Instruction-Guided mT5 Transformer Model

The rapid growth of social media has also led to a rise in abusive and harmful content, which negatively affects the online environment for users. The frequent use of offensive language and hate speech contributes to making these platforms increasingly hostile. In particular, homophobic and transphobic remarks target members of the LGBT+ community. Detecting such comments is therefore essential so that they can be flagged promptly and appropriate warnings can be given to users involved in such behaviour. The problem becomes more serious when such content appears in other forms of communication used by younger generations, such as memes. This work tries to address this issue. We propose a method to detect such content using the meme dataset from the LT-EDI 2026 challenge and secured 8th rank for English and 6th rank for Chinese language dataset in the shared task. Our approach uses a multimodal technique that processes both image and text information. The dataset has limited data, which creates a challenge. To handle this, we pre–fine-tune the models on a similar dataset called PrideMM. The proposed multimodal approach achieved Macro F1-scores of 0.24 and 0.57 for English and Chinese memes respectively

CuriousVectors@LT-EDI 2026: Detection of Homophobic and Transphobic Memes on Social Media Using a Hybrid Multimodal Approach

Online hate speech is spreading rapidly, creat ing significant challenge, particularly in low resource language such as Tamil. Lack of de veloped automated content moderation systems makes it difficult to control harmful content effectively. In this study, we propose a com putational framework for generating Counter Narratives (CNs) using classical NLP tech niques. With this, we leverage TF-IDF fea tures with n-grams to identify the labels as Ho mophobic or Transphobic. Span detection is performed with TF-IDF features with n-grams and Machine learning models. Counter narra tives are then retrieved by computing cosine similarity, ensuring semantic alignment and contextual relevance. Evaluation on the ex panded human curated dataset demonstrates that our approach produces contextually appro priate and semantically coherent counter nar ratives. Notably, the proposed system is sub mitted at Task 2 shown a overall average score of 80.40 % for Tamil and 77.29 % for English and secured first and fourth rank respectively. GitHub: https://github.com/kannanrrk/Span Counter-Feature-Based

DLRG@LT-EDI 2026: Automating Counter-Narratives for Homophobic and Transphobic Comments

The detection and response to homophobic and transphobic comments are important chal lenges in Natural Language Processing. In this paper, we focus on the detection of span for homophobic and transphobic comments (Task 1) and generation of counter narratives for abu sive comments (Task 2) for the LT-EDI @ ACL 2026 shared task. Harmful comments made online against the LGBTQ+ community have created a hostile environment for users. In this paper, we have used the transformer model for the detection of span for homophobic and trans phobic comments and generation of counter narratives. In this task, the detection of the span of comments containing homophobic and trans phobic words and the generation of counter nar ratives for abusive comments have been done using the transformer model. The results show the efficiency of the transformer model in the detection of the span of comments and gener ation of counter narratives. This paper empha sizes the efficiency of the transformer model in creating a safe environment for users.

DuoNova@LTEDI 2026: Multilingual Span Detection and Counter-Narrative Generation on Homophobic and Transphobic Comments

The deployment of Large Language Models (LLMs) has intensified concerns regarding the propagation of societal stereotypes encoded with web-scale training corpora. This pa per presents a dual-paradigm framework spe cially designed to address multilingual gender inclusivity and counterfactual generation. For multilingual gender-neutral text transformation, a fine-tuned mT5 encoder–decoder model per forms controlled sentence rewriting with mini mal edits while preserving semantic fidelity and grammatical fluency. For counter-narrative gen eration, the Llama-3 8B decoder-only model is employed to generate empathetic and persua sive responses through structured prompt-based generation. The framework is evaluated using datasets from the LT-EDI ACL 2026 shared task across multiple languages, including En glish, Tamil, Kannada, German, and Spanish. Experimental results demonstrate strong effec tiveness in identifying and neutralizing gender markers, particularly in morphologically rich languages, while the counter-narrative compo nent achieves high performance in politeness, coherence, and relevance. Overall, the pro posed approach contributes toward the develop ment of responsible and inclusive multilingual NLPsystems.

Igniters@LTEDI 2026: Multilingual Gender-Inclusive Language Generation with mT5 and Counter-Narrative Generation Using Llama-3

This paper describes the IHLC team’s submis sion to the LT-EDI 2026 shared task on gender inclusive language generation. For Subtask A(gender-neutral rewriting), we applied Low Rank Adaptation (LoRA) fine-tuning, achiev ing an 80.00% evaluation score and placing 3rd. Our primary methodological focus, how ever, is Subtask B (counter-narrative genera tion), where we propose a compute-efficient representation engineering approach. We com pute a PCA-derived steering direction from counterfactual activations and inject it into the Gemma-3-4B-it model at inference time, shift ing behavior toward inclusivity without weight updates. Paired with constrained prompting, this yielded polite, context-aware responses and a 78.12% score (Rank 6). We conclude with a manual evaluation of steering failure modes, detailing critical trade-offs in semantic preser vation and over-steering instability.

IHLC@LT-EDI 2026: Steering Toward Inclusivity - A Representation Engineering for Gender-Neutral Rewriting

This paper presents our submissions to the LT EDI@ACL 2026 Shared Task on Gender Inclu sive Language Generation. The task focuses on controlled text rewriting that reduces gen der bias while keeping the original meaning and fluency intact. We participated in both the subtasks and treated them independently, training separate instances of the instruction tuned encoder–decoder model on the respective training datasets. Scores are calculated based on averages across different rubrics, including Gender Assumption (GA), Gender Neutrality (GN), and Quality Relevance (QR) for Task A, and Politeness and Respectful (PR), Con textual Counter-Narrative Coherence (CCNC), and Quality and Relevance (QR) for Task B. For Subtask A (Gender-Inclusive Language Generation) in the English dataset, an aver age score of 43.7917 could be achieved. For Subtask B (Counterfactual Generation), we achieved an average score of 82.6241. Over all, the experiments indicate that full finetuning of instruction-tuned transformers provides an effective way to produce sentence in gender neutral form and also producing counter-factual sentences for biased one, when each subtask is optimized on its own data.

Downloads

Next from ACL 2026

From Form to Meaning: Interlingua Sense-Alignment of Offensive Language with LLMs

Similar lecture

Linguistic Corpus Annotation for Automatic Text Simplification Evaluation

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES

.css-70qvj9{display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-align-items:center;-webkit-box-align:center;-ms-flex-align:center;align-items:center;}Downloads

Next from ACL 2026

From Form to Meaning: Interlingua Sense-Alignment of Offensive Language with LLMs

Similar lecture

Linguistic Corpus Annotation for Automatic Text Simplification Evaluation

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES

Downloads