Research map / Image generation / Text-to-image diffusion
Crossing Borders: A Multimodal Challenge for Indian Poetry Translation and Image Generation
Sofia Jamil, Kotla Sai Charan, Sriparna Saha, Koustava Goswami, Joseph K J
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add ORPO alignment after fine-tuning | Gemma 2 · MorphoVerse poem translation to English | BLEU4: 0.2576 -> 0.2864 | Improved |
| add ORPO alignment after fine-tuning | Llama 3.1 · MorphoVerse poem translation to English | BLEU4: 0.2189 -> 0.2277 | Improved |
| add ORPO alignment after fine-tuning | Qwen 2.5 · MorphoVerse poem translation to English | BLEU4: 0.1674 -> 0.2099 | Improved |
| add ORPO alignment after fine-tuning | Mistral · MorphoVerse poem translation to English | BLEU4: 0.2036 -> 0.1858 | Hurt |
| add ORPO alignment after fine-tuning | Sarvam 1 · MorphoVerse poem translation to English | BLEU4: 0.1815 -> 0.2079 | Improved |
| add DPO alignment after fine-tuning | Llama 3.1 · MorphoVerse poem translation to English | BLEU4: 0.2189 -> 0.1784 | Hurt |
| add DPO alignment after fine-tuning | Gemma 2 · MorphoVerse poem translation to English | BLEU4: 0.2576 -> 0.2458 | Hurt |
| add DPO alignment after fine-tuning | Qwen 2.5 · MorphoVerse poem translation to English | BLEU4: 0.1674 -> 0.1776 | Improved |
| fine-tune on MorphoVerse instead of zero-shot | Qwen 2.5 · MorphoVerse poem translation to English | BLEU4: 0.0828 -> 0.1674 | Improved |
| fine-tune on MorphoVerse instead of zero-shot | Gemma 2 · MorphoVerse poem translation to English | BLEU4: 0.2389 -> 0.2576 | Improved |
| build image prompts from a semantic graph instead of directly prompting GPT with the translated poem (ablation) | SD 3.5 medium · MorphoVerse poem-to-image generation | Image Reward: 0.3842 -> 0.5342 | Improved |
| build image prompts from a semantic graph instead of feeding the poem directly (ablation) | SD 3.5 medium · MorphoVerse poem-to-image generation | BLIP score: 0.3813 -> 0.4613 | Improved |
Closest papers by meaning
- AltDiffusion: A Multilingual Text-to-Image Diffusion Model (AAAI 2024 · similarity 0.85)
- OpenING: A Comprehensive Benchmark for Judging Open-ended Interleaved Image-Text Generation (CVPR 2025 · similarity 0.85)
- Beyond Aesthetics: Cultural Competence in Text-to-Image Models (NeurIPS 2024 · similarity 0.84)
- PoetryDiffusion: Towards Joint Semantic and Metrical Manipulation in Poetry Generation (AAAI 2024 · similarity 0.84)
- IndicVisionBench: Benchmarking Cultural and Multilingual Understanding in VLMs (ICLR 2026 · similarity 0.84)
- MuLan: Adapting Multilingual Diffusion Models for Hundreds of Languages with Negligible Cost (ICML 2025 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
