Timezone: »

 
Poster
IGLUE: A Benchmark for Transfer Learning across Modalities, Tasks, and Languages
Emanuele Bugliarello · Fangyu Liu · Jonas Pfeiffer · Siva Reddy · Desmond Elliott · Edoardo Maria Ponti · Ivan Vulić

Wed Jul 20 03:30 PM -- 05:30 PM (PDT) @ Hall E #220

Reliable evaluation benchmarks designed for replicability and comprehensiveness have driven progress in machine learning. Due to the lack of a multilingual benchmark, however, vision-and-language research has mostly focused on English language tasks. To fill this gap, we introduce the Image-Grounded Language Understanding Evaluation benchmark. IGLUE brings together—by both aggregating pre-existing datasets and creating new ones—visual question answering, cross-modal retrieval, grounded reasoning, and grounded entailment tasks across 20 diverse languages. Our benchmark enables the evaluation of multilingual multimodal models for transfer learning, not only in a zero-shot setting, but also in newly defined few-shot learning setups. Based on the evaluation of the available state-of-the-art models, we find that translate-test transfer is superior to zero-shot transfer and that few-shot learning is hard to harness for many tasks. Moreover, downstream performance is partially explained by the amount of available unlabelled textual data for pretraining, and only weakly by the typological distance of target–source languages. We hope to encourage future research efforts in this area by releasing the benchmark to the community.

Author Information

Emanuele Bugliarello (University of Copenhagen)
Fangyu Liu (University of Cambridge)
Jonas Pfeiffer (TU-Darmstadt)
Siva Reddy (Mila)
Desmond Elliott (University of Copenhagen)
Edoardo Maria Ponti (Mila Montreal / University of Cambridge)
Ivan Vulić (University of Cambridge)

Related Events (a corresponding poster, oral, or spotlight)

More from the Same Authors