Search for: All records

We consider a hierarchical inference system with multiple clients connected to a server via a shared communication resource. When necessary, clients with low-accuracy machine learning models can offload classification tasks to a server for processing on a high-accuracy model. We propose a distributed online offloading algorithm which maximizes the accuracy subject to a shared resource utilization constraint thus indirectly realizing accuracy-delay tradeoffs possible given an underlying network scheduler. The proposed algorithm, named Lyapunov-EXP4, introduces a loss structure based on Lyapunov-drift minimization techniques to the bandits with expert advice framework. We prove that the algorithm converges to a near-optimal threshold policy on the confidence of the clients’ local inference without prior knowledge of the system’s statistics and efficiently solves a constrained bandit problem with sublinear regret. We further consider settings where clients may employ multiple thresholds, allowing more aggressive optimization of overall accuracy at a possible loss in fairness. Extensive simulation results on real and synthetic data demonstrate convergence of Lyapunov-EXP4, and show the 

  SummaryReconstructing haplotypes of an organism from a set of sequencing reads is a computationally challenging (NP-hard) problem. In reference-guided settings, at the core of haplotype assembly is the task of clustering reads according to their origin, i.e. grouping together reads that sample the same haplotype. Read length limitations and sequencing errors render this problem difficult even for diploids; the complexity of the problem grows with the ploidy of the organism. We present XHap, a novel method for haplotype assembly that aims to learn correlations between pairs of sequencing reads, including those that do not overlap but may be separated by large genomic distances, and utilize the learned correlations to assemble the haplotypes. This is accomplished by leveraging transformers, a powerful deep-learning technique that relies on the attention mechanism to discover dependencies between non-overlapping reads. Experiments on semi-experimental and real data demonstrate that the proposed method significantly outperforms state-of-the-art techniques in diploid and polyploid haplotype assembly tasks on both short and long sequencing reads. Availability and implementationThe code for XHap and the included experiments is available at https://github.com/shoryaconsul/XHap.