Fact verification pipelines are often unrealistic: they assume full-input access, operate on isolated atomic claims, and overlook the central question of when a system should act under partial evidence. We introduce DIVER (Deliberative and Iterative fact VERification), an LLM-based framework for calibrated incremental fact-checking. DIVER incrementally extracts verifiable atomic claims from progressively revealed input, retrieves and refines evidence through iterative web search, and triggers a correction as soon as a claim is reliably refuted; if no such trigger occurs, it performs a passage-level revision over the full input to recover missed errors. Across standard fact-checking benchmarks, DIVER outperforms strong LLM-based verification pipelines, with especially large gains on retrieval-intensive settings. We further introduce UmActually, a real-world benchmark built from the Internet game Um, Actually, to evaluate not only whether a system finds an error, but whether it does so at the right time. On UmActually, DIVER reports the strongest end-to-end results while typically buzzing close to the earliest valid decision point, supporting its effectiveness for calibrated incremental fact-checking.

Existing audio question answering benchmarks largely emphasize sound event classification or caption-grounded queries, often enabling models to succeed through shortcut strategies, short-duration cues, lexical priors, dataset-specific biases, or even bypassing audio via metadata and captions rather than genuine reasoning Thus, we present AUDITA (Audio Understanding from Diverse Internet Trivia Authors), a large-scale, real-world benchmark to rigorously evaluate audio reasoning beyond surface-level acoustic recognition. AUDITA comprises carefully curated, human-authored trivia questions grounded in real-world audio, designed to stress robust auditory reasoning through challenging distractors and long-range temporal dependencies, using probing queries that cannot be answered from isolated text or sound cues alone. Human average accuracy of 32.13% shows both the challenge of the task while demonstrating meaningful comprehension of the audio. In stark contrast, state-of-the- art audio question answering models perform poorly, with average accuracy below 8.86%. Beyond raw accuracy, we apply Item Response Theory (IRT) to estimate latent proficiency, question difficulty, and expose systematic deficiencies of the models and data.

Training question-answering QA and information retrieval systems for web queries require large, expensive datasets that are difficult to annotate and time-consuming to gather. Moreover, while natural datasets of information-seeking questions are often prone to ambiguity or ill-formed, there are troves of freely available, carefully crafted question datasets for many languages. Thus, we automatically generate shorter, information-seeking questions, resembling web queries in the style of the Natural Questions (NQ) dataset from longer trivia data. Training a QA system on these transformed questions is a viable strategy for alternating to more expensive training setups showing the F1 score difference of less than six points and contrasting the final systems.

This paper describes the system submitted to Large-Scale Multilingual Shared Task (Small Task #2) at WMT 2021. It is based on the massively multilingual open-source model FLORES101_MM100 model, with selective fine-tuning. Our best-performing system reported a 15.72 average BLEU score for the task.

This paper describes the UMD submission to the Explainable Quality Estimation Shared Task at the EMNLP 2021 Workshop on “Evaluation & Comparison of NLP Systems”. We participated in the word-level and sentence-level MT Quality Estimation (QE) constrained tasks for all language pairs: Estonian-English, Romanian-English, German-Chinese, and Russian-German. Our approach combines the predictions of a word-level explainer model on top of a sentence-level QE model and a sequence labeler trained on synthetic data. These models are based on pre-trained multilingual language models and do not require any word-level annotations for training, making them well suited to zero-shot settings. Our best-performing system improves over the best baseline across all metrics and language pairs, with an average gain of 0.1 in AUC, Average Precision, and Recall at Top-K score.