We present Segment Anything Model 2 (SAM 2), a foundation model towards solving promptable visual segmentation in images and videos. We build a data engine, which improves model and data via user interaction, to collect the largest video segmentation dataset to date. Our model is a simple transformer architecture with streaming memory for real-time video processing. SAM 2 trained on our data provides strong performance across a wide range of tasks. In video segmentation, we observe better accuracy, using 3x fewer interactions than prior approaches. In image segmentation, our model is more accurate and 6x faster than the Segment Anything Model (SAM). We believe that our data, model, and insights will serve as a significant milestone for video segmentation and related perception tasks. We are releasing our main model, the dataset, an interactive demo and code.

Window attention, position embeddings, and high resolution finetuning are core concepts in the modern transformer era of computer vision. However, we find that naively combining these near ubiquitous components can have a detrimental effect on performance. The issue is simple: interpolating position embeddings while using window attention is wrong. We study two state-of-the-art methods that have these three components, namely Hiera and ViTDet, and find that both do indeed suffer from this bug. To fix it, we introduce a simple absolute window position embedding strategy, which solves the bug outright in Hiera and allows us to increase both speed and performance of the model in ViTDet. We finally combine the two to obtain HieraDet, which achieves 61.7 box mAP on COCO, making it state-of-the-art for models that only use ImageNet-1k pretraining. This all stems from what is essentially a 3 line bug fix, which we name "absolute win".

Modern hierarchical vision transformers have added several vision-specific components in the pursuit of supervised classification performance. While these components lead to effective accuracies and attractive FLOP counts, the added complexity actually makes these transformers slower than their vanilla ViT counterparts. In this paper, we argue that this additional bulk is unnecessary. By pretraining with a strong visual pretext task (MAE), we can strip out all the bells-and-whistles from a state-of-the-art multi-stage vision transformer without losing accuracy. In the process, we create Hiera, an extremely simple hierarchical vision transformer that is more accurate than previous models while being significantly faster both at inference and during training. We evaluate Hiera on a variety of tasks for image and video recognition. Our code and models are available at https://github.com/facebookresearch/hiera.

We present Masked Audio-Video Learners (MAViL) to learn audio-visual representations with three complementary forms of self-supervision: (1) reconstructing masked raw audio and video inputs, (2) intra-modal and inter-modal contrastive learning with masking, and (3) self-training to predict aligned and contextualized audio-video representations learned from the first two objectives. Empirically, MAViL achieves state-of-the-art audio-video classification performance on AudioSet (53.3 mAP) and VGGSound (67.1\% accuracy), surpassing recent self-supervised models and supervised models that utilize external labeled data. Notably, pre-training with MAViL not only enhances performance in multimodal classification and retrieval tasks, but it also improves the representations of each modality in isolation, without relying on information from the other modality during uni-modal fine-tuning or inference. The code and models are available at https://github.com/facebookresearch/MAViL.

Understanding how humans and animals learn about statistical regularities in stable and volatile environments, and utilize these regularities to make predictions and decisions, is an important problem in neuroscience and psychology. Using a Bayesian modeling framework, specifically the Dynamic Belief Model (DBM), it has previously been shown that humans tend to make the {\it default} assumption that environmental statistics undergo abrupt, unsignaled changes, even when environmental statistics are actually stable. Because exact Bayesian inference in this setting, an example of switching state space models, is computationally intense, a number of approximately Bayesian and heuristic algorithms have been proposed to account for learning/prediction in the brain. Here, we examine a neurally plausible algorithm, a special case of leaky integration dynamics we denote as EXP (for exponential filtering), that is significantly simpler than all previously suggested algorithms except for the delta-learning rule, and which far outperforms the delta rule in approximating Bayesian prediction performance. We derive the theoretical relationship between DBM and EXP, and show that EXP gains computational efficiency by foregoing the representation of inferential uncertainty (as does the delta rule), but that it nevertheless achieves near-Bayesian performance due to its ability to incorporate a "persistent prior" influence unique to DBM and absent from the other algorithms. Furthermore, we show that EXP is comparable to DBM but better than all other models in reproducing human behavior in a visual search task, suggesting that human learning and prediction also incorporates an element of persistent prior. More broadly, our work demonstrates that when observations are information-poor, detecting changes or modulating the learning rate is both {\it difficult} and (thus) {\it unnecessary} for making Bayes-optimal predictions.

Understanding how humans perceive the likability of high-dimensional objects'' such as faces is an important problem in both cognitive science and AI/ML. Existing models generally assume these preferences to be fixed. However, psychologists have found human assessment of facial attractiveness to be context-dependent. Specifically, the classical Beauty-in-Averageness (BiA) effect, whereby a blended face is judged to be more attractive than the originals, is significantly diminished or reversed when the original faces are recognizable, or when the blend is mixed-race/mixed-gender and the attractiveness judgment is preceded by a race/gender categorization, respectively. This "Ugliness-in-Averageness" (UiA) effect has previously been explained via a qualitative disfluency account, which posits that the negative affect associated with the difficult race or gender categorization is inadvertently interpreted by the brain as a dislike for the face itself. In contrast, we hypothesize that human preference for an object is increased when it incurs lower encoding cost, in particular when its perceived {\it statistical typicality} is high, in consonance with Barlow's seminalefficient coding hypothesis.'' This statistical coding cost account explains both BiA, where facial blends generally have higher likelihood than ``parent faces'', and UiA, when the preceding context or task restricts face representation to a task-relevant subset of features, thus redefining statistical typicality and encoding cost within that subspace. We use simulations to show that our model provides a parsimonious, statistically grounded, and quantitative account of both BiA and UiA. We validate our model using experimental data from a gender categorization task. We also propose a novel experiment, based on model predictions, that will be able to arbitrate between the disfluency account and our statistical coding cost account of attractiveness.