Vision & Multimodal Sep 2025 · #19 most cited · 194 citations

Video models are zero-shot learners and reasoners

Thaddäus Wiedemer et al.

arXiv:2509.20328

In short

The authors test Google’s Veo 3 video model on tasks it was never trained for, such as segmentation, edge detection, image editing, physics understanding and maze solving, just by prompting it. It handles many of them, which they read as early visual reasoning.

Why it matters

It suggests video models may become general vision foundation models the way LLMs did for text.

Read first

The 3 Field Guide ideas this paper leans on.

Starting from scratch? The full route 24 ideas · basics first
  1. Dataset ✓ understood

    A collection of data examples used for training, validating, or testing machine learning models.

  2. Training ✓ understood

    The process of fitting a model to data by repeatedly measuring how wrong its outputs are and adjusting its parameters to reduce that error.

  3. Machine Learning ✓ understood

    Building systems that learn patterns from data instead of following hand-written rules, getting better at a task as they see more examples.

  4. Unsupervised Learning ✓ understood

    Learning from unlabeled data to discover hidden patterns, structures, or relationships without explicit target outputs.

  5. Self-Supervised Learning ✓ understood

    Training on unlabeled data by making the data supply its own answers: hide or alter part of an input and have the model predict or match it.

  6. Pre-training ✓ understood

    Training a model on a large dataset (often self-supervised) before fine-tuning on specific tasks, enabling transfer learning.

  7. Transfer Learning ✓ understood

    Leveraging knowledge learned from one task/domain to improve performance on a related task with less data.

  8. Foundation Model · read first ✓ understood

    A large model pre-trained on broad data at scale that can be adapted to many downstream tasks, such as GPT-style LLMs, BERT or CLIP.

  9. Zero-Shot Learning · read first ✓ understood

    A model's ability to perform tasks it wasn't explicitly trained on, using only instructions or descriptions.

  10. Natural Language Processing ✓ understood

    The field of AI that lets computers read, interpret, translate and generate human language, from spam filters and search to chatbots.

  11. Token ✓ understood

    The unit of text a language model reads and writes: a word, part of a word, or a character, drawn from the model's fixed vocabulary.

  12. Tokenization ✓ understood

    Splitting text into tokens, usually subword pieces, and mapping each to an integer ID so a language model can process it.

  13. Language Modeling ✓ understood

    Assigning probabilities to sequences of text, in practice by predicting each next token from the ones before it; the objective behind large language models.

  14. Feature ✓ understood

    A single measurable property of an example, such as a house's floor area or how many links an email contains, used as an input to a model.

  15. Neural Network ✓ understood

    A model built from layers of simple units, each taking a weighted sum of its inputs and applying a nonlinearity; training adjusts the weights so the whole stack maps inputs to useful outputs.

  16. Deep Learning ✓ understood

    A subset of machine learning that uses neural networks with multiple layers (deep neural networks) to learn hierarchical representations of data.

  17. Representation Learning ✓ understood

    Letting a model learn its own internal description of the data, such as embeddings or layered visual features, instead of relying on features designed by hand.

  18. Embedding ✓ understood

    A list of numbers (a vector) that represents a word, sentence, image or other item, learned so that similar items end up close together.

  19. Attention Mechanism ✓ understood

    A technique that lets a neural network weigh every part of its input when producing each output, focusing on the parts most relevant at that step.

  20. Transformer ✓ understood

    A neural network architecture, introduced in 2017, built from stacked self-attention and feed-forward layers; the basis of nearly every modern large language model.

  21. Large Language Model ✓ understood

    A neural network, almost always a transformer, trained on vast amounts of text to predict the next token, which lets it write, answer, summarize and follow instructions.

  22. Compute ✓ understood

    Informal term for computational resources (GPUs, TPUs, time) required for training or running AI models.

  23. Neural Scaling Laws ✓ understood

    Empirical relationships showing how model performance improves predictably with model size, data, and compute.

  24. Emergent Abilities · read first ✓ understood

    Capabilities that appear suddenly in large language models at certain scales, not present in smaller models.

In the frontier

Rank
#19 of 100
Citations
194
as of Aug 9, 2026
Published
Sep 2025

Topics: Video understanding , Reasoning methods , Vision-language models

Selection: 1kpapers.com by Together AI, most-cited as of Aug 9, 2026

Nearby papers

Summary in our own words; read the paper for the details. ← All papers