Coordination lies at the heart of many economic phenomena. A well-known example is currency crises, in which traders decide whether to launch a speculative attack. On one hand, shocks to the currency's fundamentals can propagate: as more traders attack, the central bank's foreign reserves are depleted, which in turn encourages further attacks as traders seek to exploit a weakening currency. On the other hand, shocks can also fizzle out: traders may quickly learn that the central bank's balance sheet is strong, causing pessimism to dissipate and attacks to subside. When do shocks propagate, and when do they fizzle out? In particular, how do these outcomes depend on the speed at which traders learn about the fundamental?

MOTIVATION: Single-cell DNA sequencing (scDNA-seq) and multi-modal profiling with the addition of cell-surface antibodies (scDAb-seq) have recently provided key insights into cancer heterogeneity. Scaling these technologies across large patient cohorts, however, is cost and time prohibitive. Multiplexing, in which cells from unique patients are pooled into a single experiment, offers a possible solution. While multiplexing methods exist for scRNAseq, accurate demultiplexing in scDNAseq remains an unmet need. RESULTS: Here, we introduce SNACS: single-nucleotide polymorphism and antibody-based cell sorting. SNACS relies on a combination of patient-level cell-surface identifiers and natural variation in genetic polymorphisms to demultiplex scDNAseq data. We demonstrated the performance of SNACS on a dataset consisting of multi-sample experiments from patients with leukemia where we knew truth from single-sample experiments from the same patients. Using SNACS, accuracy ranged from 0.948 to 0.991 versus 0.552 to 0.934 using demultiplexing methods from the single-cell literature. AVAILABILITY AND IMPLEMENTATION: SNACS is available at https://github.com/olshena/SNACS.

Coordination and learning both play fundamental roles in social and economic life. For instance, investors would like to coordinate their loan decisions with other investors, but also learn over time about the profitability of a company; consumers would like to adopt a networked product if other consumers do the same, but also learn over time about the quality of the product; citizens would like to protest if others also turn out, but also learn over time about the venality of the current regime. In each of these examples, whether coordination succeeds or fails is shaped by the dynamic information environment---who learns what, what is learnt, and how quickly. Thus, dynamic information is a powerful tool for shaping coordination outcomes.

In this paper, we examine the use of Transfer Learning using Pretrained Audio Neural Networks (PANNs) [1], and propose an architecture that is able to better leverage the acoustic features provided by PANNs for the Automated Audio Captioning Task [2]. We also introduce a novel self-supervised objective, Reconstruction Latent Space Similarity Regularization (RLSSR). The RLSSR module supplements the training of the model by maximizing the similarity between the encoder and decoder embedding. The combination of both methods allows us to surpass state of the art results by a significant margin on the Clotho dataset [3] across several metrics and benchmarks.

Bi-level programming problems arise in situations when the decision maker has to take into account the responses of the users to his decisions. These problems are recognized as one of the most difficult and challenging problems in transportation systems management. Several problems within the transportation literature can be cast in the bi-level programming framework. At the same time, significant advances have been made in the deployment of stochastic heuristics for function optimization. This paper reports on the use of Differential Evolution (DE) for solving bi-level programming problems with applications in the field of transportation planning. After illustrating our solution algorithm with some mathematical functions, we then apply this method to two control problems facing the transportation network manager. DE is integrated with conventional traffic assignment techniques to solve the resulting bi-level program. Numerical computations of this DE based algorithm (known as DEBLP) are presented and compared with existing results. Our numerical results augment the view that DE is a suitable contender for solving these types of problems.