What does it take for an MT system to stay semantically faithful while still landing in the right language, domain, and register? HexaBee-4B tackles that question with a three-stage recipe for six-language translation built on Gemma 3 4B IT, while PatTransBench introduces a 12,000-item Chinese-English/Japanese/Korean patent benchmark that separates sentence and paragraph cases and scores them across four dimensions. SIREN shifts the focus to crisis messaging, asking whether translations preserve protective action, directive force, and urgency, and PATMT2 reports a two-stage LoRA adaptation for patent translation that lifts a four-language Hy-MT2-7B model on six directions while checking how far the adaptation holds outside the patent domain.

HexaBee-4B: Six-Language Translation with Multilingual Answering and Bidirectional Supervision

HexaBee-4B gains chrF++ on all 30 directions, so a six-language MT stack can benefit from joint multilingual answering plus synthetic forward/reverse translation data rather than translation-only fine-tuning.

Multilingual translation requires a model to preserve meaning while producing text in a requested language. We present HexaBee-4B, a translator adapted from Gemma 3 4B IT through three stages: Wikipedia continued pretraining, joint multilingual answering and answer-text translation, and bidirectional translation specialization. The intermediate stage connects two tasks through shared multilingual answers: answering a question in the requested language and translating an answer into that language. It comprises 19.46 million answering records, including 5.76 million cross-language question-answer pairs, and 10.70 million answer-text translations. The final stage uses 658,274 forward and reverse synthetic translation examples. On 36,572 synthetic test examples spanning all 30 directions among Arabic, Chinese, English, Hebrew, Persian, and Spanish, HexaBee-4B achieves direction-macro chrF++ of 73.75, compared with 57.11 for Gemma 3 4B IT and 66.83 for Gemma 3 12B IT. BLEU, TER, COMET, and a matched model-judged subset support the same overall ranking. These results demonstrate improved translation across the six-language set, with chrF++ gains in every evaluated direction over both baselines.

PatTransBench: an ipc-stratified, multi-dimensional benchmark for patent machine translation evaluation

PatTransBench gives patent MT a 12,000-item Chinese-English/Japanese/Korean benchmark with short and long texts, so evaluate beyond BLEU because quantization can change directions and language failures in ways aggregate scores hide.

We present PatTransBench, a 12,000-item Chinese-English/Japanese/Korean patent machine translation (PatMT) benchmark with a four-dimensional evaluation protocol. The benchmark comprises 6,000 short-text (sentence/claim) and 6,000 long-text (paragraph) pairs, evenly split across three language pairs and both translation directions, built via IPC-section-stratified sampling and dualthreshold quality gating from upstream patent-family parallel corpora. The evaluation protocol spans literal n-gram overlap (BLEU, ROUGE-L, chrF), cross-lingual semantic textual similarity (CLSTS), 38-probe-language consistency testing, and serving performance under concurrency. A case study on PatMT2-7B and its fp8/AWQ4/AutoRound-4bit variants reveals direction-dependent quantization effects masked by aggregate scores and non-monotonic language-failure patterns across quantization schemes that are invisible to single-metric evaluation. We plan to release the benchmark, protocol, and case-study results upon publication to support reproducible PatMT evaluation.

SIREN: Situation-aware Index for Response and Emergency ActioNability in Crisis Translation

SIREN asks you to judge whether crisis translations preserve protective action, directive force, and urgency, and its annotations show that 31.63% of ChatGPT-4o errors and 48.73% of Google Translate errors break actionability.

Machine translation (MT) is increasingly utilized in multilingual crisis communication, yet conventional evaluation does not directly assess whether translated emergency messages preserve their actionability. We introduce SIREN (Situation-aware Index for Response and Emergency ActioNability), a human evaluation framework assessing whether translation choices compromise protective action, directive force, and urgency of the source text. We evaluate SIREN in two complementary settings. At the error level, we apply it to HOPE-annotated Italian--English translations produced by ChatGPT-4o and Google Translate: 31.63% and 48.73% of errors, respectively, compromise actionability, predominantly by affecting the protective action. At the message level, we evaluate English--French translations produced by TranslateGemma-4B and Google Translate, with actionability preserved in 85.34% and 92.67% of translations, respectively. TranslateGemma-4B failures predominantly weaken directive force, whereas Google Translate failures more often compromise protective actions and receive Major impact ratings. These findings show that translation systems can differ not only in how often actionability is lost, but also in the nature and functional impact of these failures. Alongside SIREN, we contribute an English--French crisis-translation resource with human actionability annotations and post-edited reference translations, supporting further research on function-oriented evaluation for crisis communication. Our experimental data is available at https://github.com/mcstaiano/SIREN.

PATMT2: Two-stage domain adaptation for multilingual patent translation

PatMT2’s two-stage LoRA adaptation lifts six patent directions by up to 8.8 BLEU, but probing 38 unseen languages finds four low-resource failures, so domain gains do not imply broad multilingual retention.

We present PatMT2 (Patent Machine Translation, v2), a two-stage LoRA pipeline that adapts the Hy-MT2-7B dedicated translation model to four-language (Chinese/English/Japanese/Korean) patent translation. On six translation directions, the Stage-2 checkpoint reaches a macro-average BLEU of 43.39 (BF16) to 43.57 (AWQ INT4), improving over the Hy-MT2-7B base by up to +8.8 BLEU and also exceeding Qwen3-32B and several external systems on the same average. We then measure the cost of adaptation: probing 38 languages absent from the fine-tuning data, the model still produces correct, fluent output on 34, while four low-resource languages collapse into distinct, reproducible failure modes. We treat these as diagnostic cases rather than retention rates, and report their mechanisms without attributing them to a single cause. Paired bootstrap tests confirm the translation gains over all baselines (p < 0.001).

KhyontekAI at CheckThat! 2026: Multilingual Dense Retrieval and Cross-Encoder Reranking for Scientific Web Claims

This multilingual source-retrieval system gets its best MRR@5 from metadata enrichment plus cross-encoder reranking, so practitioners building MT-adjacent search should treat metadata as a first-class feature.

Scientific claims shared through social media platforms frequently reference research findings without explicitly linking to the corresponding scientific publications. Automatically retrieving the correct scientific source for such claims is a challenging multilingual retrieval problem involving semantic matching, paraphrasing, noisy social media text, and cross-domain language variation. In this work, we present a multilingual two-stage retrieval framework for CLEF 2026 CheckThat! Task 1 on Source Retrieval for Scientific Web Claims. Our system combines fine-tuned dense retrieval models with cross-encoder reranking for improving multilingual scientific document retrieval. We evaluate multiple transformer-based dense retrieval architectures including multilingual-E5, BGE, and Contriever under different metadata configurations. We further investigate multilingual joint training, retrieval depth analysis, metadata enrichment, and language balancing strategies for improving retrieval effectiveness. Experimental results demonstrate that metadata enrichment consistently improves retrieval effectiveness, while fine-tuned cross-encoder reranking provides additional gains in retrieval precision. Our final multilingual system achieves MRR@5 scores of 0.6222 for English, 0.5112 for German, and 0.6254 for French, with an overall average MRR@5 score of 0.5862.

Papers announced by arXiv on Monday, 12 October 2026, covering submissions from 9 October to 12 October 2026.