Si Wu @ EMNLP

@si-wu.bsky.social

PhD candidate working on NLP, computational social science, and digital humanities at Northeastern University. Psycholinguistics, translation, multimodality, art, culture. I don't check here often. Email me!

I’m at #EMNLP2026 until Oct 30, then in Paris until Nov 8! Old friends and new: if you’re in either city, happy to chat about research or just grab coffee/food/drinks/dessert! I’m also on the lookout for postdoc and industry positions, so feel free to reach out about that too!

Literary works are often translated into the same language multiple times, and each translation interprets the original in its own way. How can we use these different translations to train better literary MT systems? Can synthetic references be as good as human ones?

Bild

Editors’ Choice: Computational Approaches to Cultural Heritage in the Global South: Infrastructures, Analytics, and Epistemic Justice

Computational Approaches to Cultural Heritage in the Global South: Infrastructures, Analytics, and Epistemic Justice

This special issue examines how computational and digital humanities methodologies can be mobilized from and with the Global South to preserve, interpret, and contest cultural heritage under conditions of uneven infrastructures, colonial archival violence, and epistemic inequality. Rather than treating “cultural analytics” as a narrow, model-driven practice applied to pre-existing corpora, the contributions gathered here show that, in many Global South contexts, knowledge production depends on the active, situated construction of what is conventionally treated as “data.” Datasets, collections, ontologies, and platforms emerge through socio-technical processes that are at once analytical, theoretical, and political, foregrounding that what counts as data is not given in advance but shaped by infrastructures, practices, and epistemic commitments, as well as by the “designated communities” whose knowledge practices, languages, and representational forms are (or are not) recognized within a system (Pereda et al., 2025). Our point of departure is that dominant templates for computational humanities work such as large already-digitized corpora, mature NLP pipelines for hegemonic languages, stable repositories, and clear separations between “data production” and “data analysis”, are historically specific to the Global North and, when universalized as neutral standards, misrecognize much Global South digital humanities labor as merely preparatory or “technical.” Under conditions of archival fragmentation, linguistic marginalization, and heterogeneous infrastructural ecologies, including social and community-based infrastructures, the conceptual design of collections, the negotiation of community governance, the creation of reusable data and metadata, and the iterative assembly of corpora are central forms of analysis that make particular worlds legible, contest others, and materially redistribute epistemic authority, while also resonating with broader calls to treat digital infrastructures as relational and contested sites of power.

culturalanalytics.org

How do embedding layer representations fare as predictors of reading times against other related predictors such as surprisal? Check out our paper 𝐏𝐫𝐨𝐛𝐢𝐧𝐠 𝐟𝐨𝐫 𝐑𝐞𝐚𝐝𝐢𝐧𝐠 𝐓𝐢𝐦𝐞𝐬 at #ACL2026 📍Oral Session C: Sun July 5, 16:00-17:30, San Diego Preprint: lnkd.in/eNcWZRjC

Bild