Modeling Uncertainty with Hedged Instance Embedding
This addresses uncertainty modeling in image embeddings for applications like recognition and retrieval, but it is incremental as it builds on existing metric learning and variational methods.
The paper tackles the problem of representing uncertainty in instance embeddings for ambiguous inputs like occluded or blurry images by introducing hedged instance embeddings (HIB) that model embeddings as random variables. Empirical results on a new N-digit MNIST dataset show improved performance in image matching and classification, more structured embedding spaces, and a per-exemplar uncertainty measure correlated with downstream performance.
Instance embeddings are an efficient and versatile image representation that facilitates applications like recognition, verification, retrieval, and clustering. Many metric learning methods represent the input as a single point in the embedding space. Often the distance between points is used as a proxy for match confidence. However, this can fail to represent uncertainty arising when the input is ambiguous, e.g., due to occlusion or blurriness. This work addresses this issue and explicitly models the uncertainty by hedging the location of each input in the embedding space. We introduce the hedged instance embedding (HIB) in which embeddings are modeled as random variables and the model is trained under the variational information bottleneck principle. Empirical results on our new N-digit MNIST dataset show that our method leads to the desired behavior of hedging its bets across the embedding space upon encountering ambiguous inputs. This results in improved performance for image matching and classification tasks, more structure in the learned embedding space, and an ability to compute a per-exemplar uncertainty measure that is correlated with downstream performance.