Research map / Vision-language and multimodal / Cross-modal learning
Jointly Discovering Visual Objects and Spoken Words from Raw Sensory Input
David Harwath, Adria Recasens, Didac Suris, Galen Chuang, Antonio Torralba, James Glass
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 |
|---|---|---|---|
| use max over image regions (MISA) instead of average (SISA) (ablation) | matchmap speech model, random init · Places val 1k caption-to-image | R@10: .274 -> .314 | Improved |
| use MISA instead of SIMA (ablation) | matchmap speech model, random init · Places val 1k caption-to-image | R@10: .284 -> .314 | Improved |
| use SIMA instead of MISA for image-to-caption (ablation) | matchmap speech model, random init · Places val 1k image-to-caption | R@1: .057 -> .065 | Mixed |
| pretrain image branch on ImageNet (ablation) | MISA speech model · Places val 1k caption-to-image | R@10: .314 -> .604 | Improved |
| pretrain image branch on ImageNet (ablation) | MISA speech model · Places val 1k image-to-caption | R@10: .291 -> .528 | Improved |
| pretrain image branch on ImageNet (ablation) | MISA speech model · ADE20k speech-prompted localization | IoU: .1795 -> .2324 | Improved |
| use MISA instead of average similarity for localization (ablation) | speech model, random init · ADE20k speech-prompted localization | IoU: .1637 -> .1795 | Improved |
| use SIMA instead of MISA for localization (ablation) | speech model, pretrained · ADE20k speech-prompted localization | IoU: .2324 -> .1857 | Hurt |
| replace ASR text input with raw speech | MISA pretrained · Places val 1k caption-to-image | R@10: .701 -> .604 | Hurt |
| replace ASR text input with raw speech | MISA pretrained · ADE20k speech-prompted localization | IoU: .2413 -> .2324 | Hurt |
Closest papers by meaning
- Unsupervised Learning of Spoken Language with Visual Context (NeurIPS 2016 · similarity 0.90)
- Separating the "Chirp" from the "Chat": Self-supervised Visual Grounding of Sound and Language (CVPR 2024 · similarity 0.88)
- Prompting Segmentation with Sound Is Generalizable Audio-Visual Source Localizer (AAAI 2024 · similarity 0.87)
- Seeing Speech and Sound: Distinguishing and Locating Audio Sources in Visual Scenes (CVPR 2025 · similarity 0.87)
- Robust Audio-Visual Segmentation via Audio-Guided Visual Convergent Alignment (CVPR 2025 · similarity 0.86)
- Hear, Localize, and Reason: Spatially Aware Scene Understanding for Audio-visual LLMs (NeurIPS 2026 · similarity 0.86)
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.
