Welcome!
I am Lateef Adeleke, a PhD candidate in Linguistics at the University of Rochester.
My work is at the intersection of language documentation, grammatical analysis, corpus linguistics, and speech technology.
I study how underdescribed African languages can contribute to linguistic theory and the design of language technologies when they are analyzed through fieldwork data, naturalistic speech, and community-based documentation.
📄 CV (PDF)
🏛️ Google Scholar ·
🐙 GitHub
✉️ ladeleke@ur.rochester.edu
Research
My work focuses on African languages, and I combine fieldwork, formal grammatical analysis, and neural modeling to build computational systems grounded in real linguistic structure.
I am currently leading a multi-year ELDP-sponsored documentation project on Uneme, an underdescribed Nigerian language. Through this project, I am developing a multimodal corpus, annotation, grammatical description, and an Uneme-English bilingual dictionary. I am also training an ASR model for Uneme, and working on the influence of genealogical relatedness, phonological similarity, tone systems, and domain variation on cross-lingual transfer in Automatic Speech Recognition (ASR).
Research Approach
My work integrates:
- Field linguistics and corpus creation I collect, annotate, and archive naturalistic and elicited language data through community-centred fieldwork.
- Variation-aware grammatical analysis I examine how grammar varies across speech varieties, with special attention to TAM, negation, focus, tone, and clause structure.
- Corpus-accountable grammar writing I use documentary corpora as evidence for grammatical description, typology, and theory-building.
- Low-resource ASR and speech technology I build and evaluate ASR systems from fieldwork data, with a special focus on the challenges posed by naturalistic speech, style mismatch, dialectal variation, and tonal contrasts.
- Linguistically-informed AI evaluation I study model errors in terms of linguistic structure by investigating how speech systems handle tone, morphology, phonological contrasts, and grammatical markers in underdescribed languages.
Broader Vision
African languages should not remain underdescribed or passive “low-resource benchmarks”. They should actively participate in the theory of language and design of next-generation language technologies. My broader goal is to develop research models where documentation, formal grammatical analysis, and computational methods work together.
News
- 2026 — At the 55th CALL in Leiden, I will give a talk on the “Reflexes of Movement in Uneme”
- 2026 — I gave a talk on “Polarity-Conditioned asymmetries in Uneme” At ACAL57 in Buffalo
- 2026 — I gave a talk on “Reflexes of Movement in Uhami” At ACAL57 in Buffalo
- 2026 — Presented “Linguistically-informed Evaluation of Multilingual ASR in African Languages” at the 7th AfricaNLP in Rabat
- 2026 — Attended the 19th Conference of the European Chapter of the Association for Computational Linguistics, in Rabat
- 2026 — Began the second round of data collection for documenting Uneme
- Sept 2025 — Presented at the University of Rochester’s Omnibus Meeting.
- June 2025 — Began data collection for Uneme documentation.
- June 2025 — Organized training on documentation methods in Nigeria.
- June 2025 — Awarded the Donald M. and Janet C. Barnard Fellowship
- May 2025 — Attended ELDP’s Language Documentation Training in Berlin.
- May 2025 — Presented at the 56th ACAL, University of Minnesota.
- April 2025 — Awarded the ELDP’s Individual Graduate Scholarship to Document Uneme Iron Technology.
Current Projects
- Creating an annotated speech corpus for Uneme
- Building a multimedia FLEx Dictionary for Uneme
Field
