Search for: All records

We fabricate three-terminal hybrid devices consisting of a semiconductor nanowire segment proximitized by a grounded superconductor and having tunnel probe contacts on both sides. By performing simultaneous tunneling measurements, we identify delocalized states, which can be observed from both ends, and states localized near one of the tunnel barriers. The delocalized states can be traced from zero magnetic field to fields beyond 0.5 T. Within the regime that supports delocalized states, we search for correlated low-energy features consistent with the presence of Majorana zero modes. While both sides of the device exhibit ubiquitous low-energy features at high fields, no correlation is inferred. Simulations using a one-dimensional effective model suggest that the delocalized states, which extend throughout the whole system, have large characteristic wave vectors, while the lower momentum states expected to give rise to Majorana physics are localized by disorder. To avoid such localization and realize Majorana zero modes, disorder needs to be reduced significantly. We propose a method for estimating the disorder strength based on analyzing the level spacing between delocalized states.

 

Latent space Energy-Based Models (EBMs), also known as energy-based priors, have drawn growing interests in generative modeling. Fueled by its flexibility in the formulation and strong modeling power of the latent space, recent works built upon it have made interesting attempts aiming at the interpretability of text modeling. However, latent space EBMs also inherit some flaws from EBMs in data space; the degenerate MCMC sampling quality in practice can lead to poor generation quality and instability in training, especially on data with complex latent structures. Inspired by the recent efforts that leverage diffusion recovery likelihood learning as a cure for the sampling issue, we introduce a novel symbiosis between the diffusion models and latent space EBMs in a variational learning framework, coined as the latent diffusion energy-based model. We develop a geometric clustering-based regularization jointly with the information bottleneck to further improve the quality of the learned latent space. Experiments on several challenging tasks demonstrate the superior performance of our model on interpretable text modeling over strong counterparts. 

We present Deep Region Competition (DRC), an algorithm designed to extract foreground objects from images in a fully unsupervised manner. Foreground extraction can be viewed as a special case of generic image segmentation that focuses on identifying and disentangling objects from the background. In this work, we rethink the foreground extraction by reconciling energy-based prior with generative image modeling in the form of Mixture of Experts (MoE), where we further introduce the learned pixel re-assignment as the essential inductive bias to capture the regularities of background regions. With this modeling, the foreground/background partition can be naturally found through Expectation-Maximization (EM). We show that the proposed method effectively exploits the interaction between the mixture components during the partitioning process, which closely connects to region competition, a seminal approach for generic image segmentation. Experiments demonstrate that DRC exhibits more competitive performances on complex real-world data and challenging multi-object scenes compared with prior methods. Moreover, we show empirically that DRC can potentially generalize to novel foreground objects even from categories unseen during training.