Papers by Manuel Mager

12 papers
Rethinking LLM Uncertainty: A Multi-Agent Approach to Estimating Black-Box Model Uncertainty (2025.findings-emnlp)

Copied to clipboard

Challenge: Existing methods to gauge model’s uncertainty through self-consistency in responses to the target query are misleading: an LLM may confidently provide an incorrect answer to a target query, yet give a confident and accurate answer to that same query when answering a knowledge-preserving perturbation of the query.
Approach: They propose a method that uses multi-agent interaction to estimate black-box LLMs' uncertainty.
Outcome: The proposed method outperforms existing self-consistency based methods and improves hallucination detection.
Exploring Segmentation Approaches for Neural Machine Translation of Code-Switched Egyptian Arabic-English Text (2023.eacl-main)

Copied to clipboard

Challenge: Code-switching (CS) is a problem in machine translation, but its performance is not investigated for CS settings.
Approach: They propose to use morphological segmentation techniques for machine translation tasks . they compare morphology-based and frequency-based segmentation for MT tasks based on data size .
Outcome: The proposed approach performs best in MT tasks but under-performs in other languages.
Large Language Models Are Overconfident in Their Own Responses (2026.findings-acl)

Copied to clipboard

Challenge: Prior work has shown that instruction-tuned large language models are less well calibrated than their base pre-trained counterparts.
Approach: They propose a simple inference-time strategy that frams the model’s answer as user input during confidence elicitation.
Outcome: The proposed approach reduces overconfidence and improves calibration by up to 26% without retraining.
Subword-Level Language Identification for Intra-Word Code-Switching (N19-1)

Copied to clipboard

Challenge: Code-switching (CS) is a phenomenon of alternating between two or more languages in conversations . if at least one language is morphologically rich, a large number of words can be composed of morphemes from more than one language.
Approach: They propose to extend the language identification task to the subword level by splitting mixed words while tagging each part with a language ID.
Outcome: The proposed model outperforms the baseline on a Spanish–Wixarika and adapted German–Turkish datasets.
AmericasNLI: Evaluating Zero-shot Natural Language Understanding of Pretrained Multilingual Models in Truly Low-resource Languages (2022.acl-long)

Copied to clipboard

Challenge: Pretrained multilingual models can perform cross-lingual transfer in a zero-shot setting, even for unseen languages.
Approach: They propose to extend XNLI to 10 indigenous languages of the Americas and test multiple zero-shot and translation-based approaches.
Outcome: The proposed model can perform cross-lingual transfer in a zero-shot setting even for languages unseen during pretraining.
Ethical Considerations for Machine Translation of Indigenous Languages: Giving a Voice to the Speakers (2023.acl-long)

Copied to clipboard

Challenge: In recent years, machine translation has become very successful for high-resource language pairs.
Approach: They conduct interviews with community leaders, teachers, and language activists to shed light on ethical considerations for the automatic translation of Indigenous languages.
Outcome: The results show that the inclusion of native speakers and community members is vital to performing better and more ethical research on Indigenous languages.
Tackling the Low-resource Challenge for Canonical Segmentation (2020.emnlp-main)

Copied to clipboard

Challenge: morphological segmentation is a task of dividing words into their constituting morphemes . we compare two new approaches for the task when training data is limited .
Approach: They propose to use an LSTM pointer-generator and a sequence-to-sequence model to perform canonical segmentation when training data is limited.
Outcome: The proposed models outperform existing models on German, English, and Indonesian in low-resource scenarios by 11.4% accuracy.
MEAV: Model Editing with Alignment Vectors for inference time LLM alignment in single and multidomain preference spectrum (2026.findings-acl)

Copied to clipboard

Challenge: Existing training-time alignment methods require full retraining when a change is needed.
Approach: They propose an inference-time model-editing-based alignment method that learns encoded representations of preference dimensions and allows dynamic adjusting of the model behavior.
Outcome: The proposed method can be used to align large language models to human preferences . it reduces the cost of inference by half compared to the prompt engineering approach .
From If-Statements to ML Pipelines: Revisiting Bias in Code-Generation (2026.findings-acl)

Copied to clipboard

Challenge: Existing methods to evaluate code generation bias focus on overt discrimination through simple conditional statements.
Approach: They examine ML pipelines that exhibit substantially greater bias than simple conditionals . they challenge simple conditional statements as valid proxies for bias evaluation .
Outcome: The proposed model underestimates real-world bias in generating machine learning pipelines . the model maintains equal performance on simple conditionals and ML pipelines, the study shows .
BPE vs. Morphological Segmentation: A Case Study on Machine Translation of Four Polysynthetic Languages (2022.findings-acl)

Copied to clipboard

Challenge: Morphologically rich polysynthetic languages present a challenge for NLP systems due to data sparsity.
Approach: They propose to use subword segmentation to reduce data sparsity in polysynthetic languages . they compare supervised and unsupervised morphological segmentation methods to Byte-Pair Encodings .
Outcome: The proposed methods outperform BPEs in MT tasks for all language pairs except for Nahuatl . the proposed methods are more efficient than supervised methods, but less sparse in fusional languages.
GPT-too: A Language-Model-First Approach for AMR-to-Text Generation (2020.acl-main)

Copied to clipboard

Challenge: Existing approaches to generating text from AMRs focus on training sequence-to-sequence or graph-tosequent models on annotated data.
Approach: They propose a strong pre-trained language model with cycle consistency-based re-scoring to generate AMR text.
Outcome: The proposed model outperforms existing methods on the English LDC2017T10 dataset.
Challenges of language technologies for the indigenous languages of the Americas (C18-1)

Copied to clipboard

Challenge: Indigenous languages of the American continent are highly diverse, but have received little attention from the technological perspective.
Approach: They review the research, the digital resources and the available NLP systems for indigenous languages of the American continent . they stress the need of developing language resources and NLP tools for these languages .
Outcome: The authors review the research and the available NLP systems on indigenous languages of the Americas . they argue that the lack of resources and tools can have a negative impact on the communities which depend on these languages .

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations