André F. T. Martins

Artificial Intelligence · with Lavie · 2022

A researcher in the Science corpus: 2 papers; trusts 5 researchers and is trusted by 2 researchers here, as inferred from the citation graph.

Subfields

• Artificial Intelligence — 40 citations into it; #32,384 of 35,345 by trust here, #23,731 by citations
• Natural Language Processing — 34 citations into it
• Evaluation & Benchmarks — 33 citations into it

Whom André F. T. Martins trusts

Their strongest outbound trust, strongest first, inferred from their citations.

1. Chi-kiu (羅致翹) Lo (in Evaluation & Benchmarks)
2. Maja Popovic (in Artificial Intelligence)
3. Ondrej Bojar (in Evaluation & Benchmarks)
4. Matt Post (in Artificial Intelligence)
5. Yvette Graham (in Evaluation & Benchmarks)

Who trusts André F. T. Martins

1. Tom Kocmi (in Artificial Intelligence)
2. Christian Federmann (in Artificial Intelligence)

Papers from André F. T. Martins's perspective

Ordered by how strongly the researchers André F. T. Martins trusts endorse each paper. Same corpus, a different order for every researcher.

1. Bleu: a Method for Automatic Evaluation of Machine Translation
2. A Call for Clarity in Reporting BLEU Scores
3. SentencePiece: A simple and language independent subword tokenizer and detokenizer for Neural Text Processing
4. BLEURT: Learning Robust Metrics for Text Generation
5. BERTScore: Evaluating Text Generation with BERT
6. Tangled up in BLEU: Reevaluating the Evaluation of Automatic Machine Translation Evaluation Metrics
7. Results of the WMT17 Metrics Shared Task
8. Results of the WMT18 Metrics Shared Task: Both characters and embeddings achieve good performance
9. Results of the WMT20 Metrics Shared Task
10. The Curious Case of Neural Text Degeneration

Their papers

• Results of WMT22 Metrics Shared Task: Stop Using BLEU – Neural Metrics Are Better and More Robust (2022) — 351 citations
• COMET-22: Unbabel-IST 2022 Submission for the Metrics Shared Task (2022) — 503 citations