Salvato in:
| Autore principale: | |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2511.18359 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866911570980765696 |
|---|---|
| author | Stergiou, Alexandros |
| author_facet | Stergiou, Alexandros |
| contents | How do video understanding models acquire their answers? Although current Vision Language Models (VLMs) reason over complex scenes with diverse objects, action performances, and scene dynamics, understanding and controlling their internal processes remains an open challenge. Motivated by recent advancements in text-to-video (T2V) generative models, this paper introduces a logits-to-video (L2V) task alongside a model-independent approach, TRANSPORTER, to generate videos that capture the underlying rules behind VLMs' predictions. Given the high-visual-fidelity produced by T2V models, TRANSPORTER learns an optimal transport coupling to VLM's high-semantic embedding spaces. In turn, logit scores define embedding directions for conditional video generation. TRANSPORTER generates videos that reflect caption changes over diverse object attributes, action adverbs, and scene context. Quantitative and qualitative evaluations across VLMs demonstrate that L2V can provide a fidelity-rich, novel direction for model interpretability that has not been previously explored. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_18359 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | TRANSPORTER: Transferring Visual Semantics from VLM Manifolds Stergiou, Alexandros Computer Vision and Pattern Recognition How do video understanding models acquire their answers? Although current Vision Language Models (VLMs) reason over complex scenes with diverse objects, action performances, and scene dynamics, understanding and controlling their internal processes remains an open challenge. Motivated by recent advancements in text-to-video (T2V) generative models, this paper introduces a logits-to-video (L2V) task alongside a model-independent approach, TRANSPORTER, to generate videos that capture the underlying rules behind VLMs' predictions. Given the high-visual-fidelity produced by T2V models, TRANSPORTER learns an optimal transport coupling to VLM's high-semantic embedding spaces. In turn, logit scores define embedding directions for conditional video generation. TRANSPORTER generates videos that reflect caption changes over diverse object attributes, action adverbs, and scene context. Quantitative and qualitative evaluations across VLMs demonstrate that L2V can provide a fidelity-rich, novel direction for model interpretability that has not been previously explored. |
| title | TRANSPORTER: Transferring Visual Semantics from VLM Manifolds |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2511.18359 |