49 citations · 51 across the 17 of their papers we have counts for
19 papers
KilometerVision: A New Frontier for Large-Scale Spatial Intelligence in VLMs
Aravindh Mahendran, Michael King, Matthew Koichi Grimes +12
We push the frontier of large-scale spatial intelligence in Vision-Language Models (VLMs) and introduce the first benchmark that probes geographical layout understanding from real-…
TAPVid-MV: A Benchmark for Tracking Any Point in 3D Across Multiple Views
Skanda Koppula, Frano Rajic, Abdullah Faiz Ur Rahman +9
Multi-camera systems are increasingly practical for robotics, AR/VR, and autonomous driving because complementary views reduce depth ambiguity and preserve visibility under occlusi…
Video Generation Models are General-Purpose Vision Learners
Letian Wang, Chuhan Zhang, Rishabh Kabra +9
Driven by next-token prediction, NLP shifted from task-specific models into powerful generalist foundation models. What, then, is the equivalent catalyst needed to achieve a genera…
Perception Test 2025: Challenge Summary and a Unified VQA Extension
Joseph Heyward, Nikhil Parthasarathy, Tyler Zhu +5
The Third Perception Test challenge was organised as a full-day workshop alongside the IEEE/CVF International Conference on Computer Vision (ICCV) 2025. Its primary goal is to benc…
Recurrent Video Masked Autoencoders
Daniel Zoran, Nikhil Parthasarathy, Yi Yang +3
We present Recurrent Video Masked-Autoencoders (RVM): a novel approach to video representation learning that leverages recurrent computation to model the temporal structure of vide…
Unique Lives, Shared World: Learning from Single-Life Videos
Tengda Han, Sayna Ebrahimi, Dilara Gokay +8
We introduce the "single-life" learning paradigm, where we train a distinct vision model exclusively on egocentric videos captured by one individual. We leverage the multiple viewp…
LayerLock: Non-collapsing Representation Learning with Progressive Freezing
Goker Erdogan, Nikhil Parthasarathy, Catalin Ionescu +5
We introduce LayerLock, a simple yet effective approach for self-supervised visual representation learning, that gradually transitions from pixel to latent prediction through progr…
Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities
Gheorghe Comanici, Eric Bieber, Mike Schaekermann +3431
In this report, we introduce the Gemini 2.X model family: Gemini 2.5 Pro and Gemini 2.5 Flash, as well as our earlier Gemini 2.0 Flash and Flash-Lite models. Gemini 2.5 Pro is our…
Learning from Streaming Video with Orthogonal Gradients
Tengda Han, Dilara Gokay, Joseph Heyward +6
We address the challenge of representation learning from a continuous stream of video as input, in a self-supervised manner. This differs from the standard approaches to video lear…
Scaling 4D Representations
João Carreira, Dilara Gokay, Michael King +32
Scaling has not yet been convincingly demonstrated for pure self-supervised learning from video. However, prior work has focused evaluations on semantic-related tasks $\unicode{x20…
TRecViT: A Recurrent Video Transformer
Viorica Pătrăucean, Xu Owen He, Joseph Heyward +10
We propose a novel block for \emph{causal} video modelling. It relies on a time-space-channel factorisation with dedicated blocks for each dimension: gated linear recurrent units (…
Perception Test 2024: Challenge Summary and a Novel Hour-Long VideoQA Benchmark
Joseph Heyward, João Carreira, Dima Damen +2
Following the successful 2023 edition, we organised the Second Perception Test challenge as a half-day workshop alongside the IEEE/CVF European Conference on Computer Vision (ECCV)…
Moving Off-the-Grid: Scene-Grounded Video Representations
Sjoerd van Steenkiste, Daniel Zoran, Yi Yang +13
Current vision models typically maintain a fixed correspondence between their representation structure and image space. Each layer comprises a set of tokens arranged "on-the-grid,"…
TAPVid-3D: A Benchmark for Tracking Any Point in 3D
Skanda Koppula, Ignacio Rocco, Yi Yang +5
We introduce a new benchmark, TAPVid-3D, for evaluating the task of long-range Tracking Any Point in 3D (TAP-3D). While point tracking in two dimensions (TAP) has many benchmarks m…
BootsTAP: Bootstrapped Training for Tracking-Any-Point
Carl Doersch, Pauline Luc, Yi Yang +8
To endow models with greater understanding of physics and motion, it is useful to enable them to perceive how solid surfaces move and deform in real scenes. This can be formalized…
Perception Test 2023: A Summary of the First Challenge And Outcome
Joseph Heyward, João Carreira, Dima Damen +2
The First Perception Test challenge was held as a half-day workshop alongside the IEEE/CVF International Conference on Computer Vision (ICCV) 2023, with the goal of benchmarking st…
Learning from One Continuous Video Stream
João Carreira, Michael King, Viorica Pătrăucean +9
We introduce a framework for online learning from a single continuous video stream -- the way people and animals learn, without mini-batches, data augmentation or shuffling. This p…
Is ImageNet worth 1 video? Learning strong image encoders from 1 long unlabelled video
Shashanka Venkataramanan, Mamshad Nayeem Rizve, João Carreira +2
Self-supervised learning has unlocked the potential of scaling up pretraining to billions of images, since annotation is unnecessary. But are we making the best use of data? How mo…
Self-supervised video pretraining yields robust and more human-aligned visual representations
Nikhil Parthasarathy, S. M. Ali Eslami, João Carreira +1
Humans learn powerful representations of objects and scenes by observing how they evolve over time. Yet, outside of specific tasks that require explicit temporal understanding, sta…