Understanding cultural heritage artifacts such as ancient Greek pottery requires expert-level reasoning that remains challenging for current MLLMs due to limited domain-specific data. We introduce VaseVQA, a benchmark for ancient Greek pottery, primarily vases, consisting of 31,773 images and 67,614 question–answer pairs across seven expert-defined categories, enabling systematic evaluation of expert-level cultural heritage understanding. Using this dataset, we explore effective training strategies for domain-specific reasoning. While supervised fine-tuning improves adaptation to domain knowledge, it struggles with deeper reasoning tasks. We propose VaseVL, which augments SFT with reinforcement learning using verifiable rewards. Experiments show that VaseVL consistently outperforms supervised baselines, especially on reasoning-intensive questions, highlighting the value of targeted reinforcement learning for cultural heritage visual question answering.

An intonational inventory of Urdu for spontaneous conversational speech is determined based on the analysis of a hand-labelled data set of telephone conversations. An inventory of Urdu pitch accents and the basic Urdu intonation patterns observed in the data are summarised and presented using a simplified version of the Rhythm and Pitch (RaP) labelling system. The relation between pitch accents and parts of speech (PoS) is also explored. The data confirm the important role played by low pitch accents in Urdu spontaneous speech, in line with previous studies on Urdu/Hindi scripted speech. Typical pitch contours such as falling tone in statements and WH-questions, and rising tone for yes/no questions are also exhibited. Pitch accent distribution is quite free in Urdu, but the data indicate a stronger association of pitch accent with some PoS categories of content word (e.g. Nouns) when compared with function words and semantically lighter PoS categories (such as Light Verbs). Contrastive focus is realised by an L*+H accent with a relatively large pitch excursion for the +H tone, and longer duration of the stressed syllable. The data suggest that post-focus compression (PFC) is used in Urdu as a focus-marking strategy.