arXiv research

End-to-end PGL framework tackles open-set domain shift.

The aim of unsupervised domain adaptation is to leverage the knowledge in a labeled (source) domain to improve a model's learning performance with an unlabeled (target) domain -- the basic strategy being to mitigate the effects of discrepancies between the two distributions. Most existing algorithms can only handle uns…

Paper tackles open set domain adaptation by detecting unknown classes.

A new method trains deep neural networks for open set domain adaptation without negative open set difference.

In conventional domain adaptation, a critical assumption is that there exists a fully labeled domain (source) that contains the same label space as another unlabeled or scarcely labeled domain (target). However, in the real world, there often exist application scenarios in which both domains are partially labeled and n…

Optimal transport method rejects new classes and adjusts class ratios for open set domain adaptation.

New method reduces label and data shifts between domains using optimal transport.

Open set recognition problems exist in many domains. For example in security, new malware classes emerge regularly; therefore malware classification systems need to identify instances from unknown classes in addition to discriminating between known classes. In this paper we present a neural network based representation…

New bounds on inscribed triangles in arbitrary planar domains.

Integrates multiple datasets to solve open set crowdsourcing problems.

Paper develops new algorithms for unsupervised multi-class domain adaptation.

Domain adaptation in person re-identification (re-ID) has always been a challenging task. In this work, we explore how to harness the natural similar characteristics existing in the samples from the target domain for learning to conduct person re-ID in an unsupervised manner. Concretely, we propose a Self-similarity Gr…

Random forest can be adapted for open-set recognition with improved performance.

Proposes IFCDA framework to improve cross-domain adaptation.

Plex improves model reliability across vision and language tasks.

For $p\in (1,2]$ and a bounded, convex, nonempty, open set $Ω\subset\mathbb R^2$ let $μ_p(\barΩ,\cdot)$ be the $p$-capacitary curvature measure (generated by the closure $\barΩ$ of $Ω$) on the unit circle $\mathbb S^1$. This paper shows that such a problem of prescribing $μ_p$ on a planar convex domain: "Given a finite…

We investigate, for the Laplacian operator, the existence and nonexistence of eigenfunctions of eigenvalue between zero and the first eigenvalue of the hyperbolic space H^n, for unbounded domains of H^n. If a domain is contained in a horoball, we prove that there is no positive bounded eigenfunction that vanishes on th…

GMVAE improves open-set classification by clustering latent representations.

Most transport theorems---that is, a formula for the rate of change of an integral in which both the integrand and domain of integration depend on time---involve domains that evolve according to a flow map. Such domains are said to be convecting. Here a transport theorem for nonconvecting domains evolving on an embedde…

Paper proposes a loss extension for neural networks to improve OSR performance.

In this paper we are interested in possible extensions of an inequality due to Minkowski: $\int_{\partialΩ} H\,dA \geq \sqrt{4πA(\partialΩ)}$ valid for any regular open set $Ω\subset\mathbb{R}^3$, where $H$ denotes the scalar mean curvature and $A$ the area. We prove that this inequality holds true for axisymmetric dom…

In this note, we investigate upper bounds of the Neumann eigenvalue problem for the Laplacian of a bounded domain (with smooth boundary) in a given complete (not compact a priori) Riemannian manifold with Ricci bounded below . For this, we use test functions for the Rayleigh quotient subordinated to a family of open se…

This paper proposes a method to use deep neural networks as end-to-end open-set classifiers. It is based on intra-class data splitting. In open-set recognition, only samples from a limited number of known classes are available for training. During inference, an open-set classifier must reject samples from unknown class…

Research improves open-set learning by leveraging unlabelled data.

Sharp stability of Alexandrov's theorem for $C^1$ domains in the small-excess regime

CGDL improves open set recognition by learning conditional Gaussian distributions.

Proposes a new framework for open set recognition using conditional probabilistic generative models.

Camera model identification refers to the problem of linking a picture to the camera model used to shoot it. As this might be an enabling factor in different forensic applications to single out possible suspects (e.g., detecting the author of child abuse or terrorist propaganda material), many accurate camera model att…

Proves a generalized isoperimetric inequality for spheres in dimensions 4 and above.

Manifold calculus of functors, due to M. Weiss, studies contravariant functors from the poset of open subsets of a smooth manifold to topological spaces. We introduce "multivariable" manifold calculus of functors which is a generalization of this theory to functors whose domain is a product of categories of open sets. …

CoDAG combines domain adaptation and generalization for unsupervised continual domain shift learning.

Proposes first method for continuously indexed domain adaptation.

This paper reviews information theory in open-world machine learning.

Survey on deep learning for malware classification, including unknown threats.

Method generates intermediate domains to align source and target domains.

The paper improves generalization bounds for domain adaptation.

Paper presents a Transformer model for automatic domain adaptation.

Dynamic residual adapters improve performance across multiple latent domains without domain labels.

In open set learning, a model must be able to generalize to novel classes when it encounters a sample that does not belong to any of the classes it has seen before. Open set learning poses a realistic learning scenario that is receiving growing attention. Existing studies on open set learning mainly focused on detectin…

Learn to automatically plug domain-specific modules into a common network.

We prove that a bounded open set U in Euclidean n-space has k-width less than C(n) Volume(U)^{k/n}. Using this estimate, we give lower bounds for the k-dilation of degree 1 maps between certain domains in Euclidean space. In particular, we estimate the smallest (n-1)-dilation of any degree 1 map between two n-dimension…

New method for estimating class proportions in open-set label shift data.

A new method uses normalizing flows for gradual domain adaptation.

We present an analysis of predictive uncertainty based out-of-distribution detection for different approaches to estimate various models' epistemic uncertainty and contrast it with extreme value theory based open set recognition. While the former alone does not seem to be enough to overcome this challenge, we demonstra…

Graph-Relational Domain Adaptation (GRDA) adapts domains based on their graph structure.

A cost-effective framework for gradual domain adaptation using multifidelity.

New method improves OSSL by learning from all unlabeled data.

Often, when dealing with real-world recognition problems, we do not need, and often cannot have, knowledge of the entire set of possible classes that might appear during operational testing. In such cases, we need to think of robust classification methods able to deal with the "unknown" and properly reject samples belo…