Neutrophils, an essential innate immune cell type with a short lifespan, rely on continuous replenishment from bone marrow (BM) precursors. Although it is established that neutrophils are derived from the granulocyte-macrophage progenitor (GMP), the molecular regulators involved in the differentiation process remain poorly understood. Here we developed a random forest-based machine-learning pipeline, NeuRGI (Neutrophil Regulatory Gene Identifier), which utilized Positive-Unlabeled Learning (PU-learning) and neural network-based in silico gene knockout to identify neutrophil regulators. We interrogated features including gene expression dynamics, physiological characteristics, pathological relatedness, and gene conservation for the model training. Our identified pipeline leads to identifying Mitogen-Activated Protein Kinase-4 (MAP4K4) as a novel neutrophil differentiation regulator. The loss of MAP4K4 in hematopoietic stem cells and progenitors in mice induced neutropenia and impeded the differentiation of neutrophils in the bone marrow. By modulating the phosphorylation level of proteins involved in cell apoptosis, such as STAT5A, MAP4K4 delicately regulates cell apoptosis during the process of neutrophil differentiation. Our work presents a novel regulatory mechanism in neutrophil differentiation and provides a robust prediction model that can be applied to other cellular differentiation processes.

We investigate the resource allocation problem, including time slot allocation, channel allocation, and power adaptation, in a millimeter Wave (mmWave) network with multiple transmission links, multiple channels, and a PicoNet Coordinator (PNC). Each link has a video session to transmit from the transmitter to the receiver. The objective is to minimize the number of time slots to finish the video sessions of all links by jointly optimizing channel allocation and time slot allocation for links, while considering the possible interference between different links on the same channel. The optimal solution for the formulated problem is computationally prohibitive to obtain due to the exponential complexity. We developed a column generation based method to reformulate the original problem into a main problem along with a series of sub-problems, with greatly reduced complexity. We prove that the optimal solution for the reformulated problem converges to the optimal solution of the original problem, and we derived a lower bound for the performance of the reformulated problem at each iteration, which will finally converge to the global optimal solution. The proposed scheme is validated with simulations with its superior performance over existing work is observed.

We investigate the scheduling problem in a centralized dual-hop 60 GHz network with multiple Source-Destination (SD) pairs, relays, and a PicoNet Coordinator (PNC). The objective is to minimize the Maximum Expected Delivery Time (MEDT) among all SD pairs by jointly optimizing relay and link selection, while exploiting reflected mmWave transmissions and considering link blockage dynamics. We develop a Decomposition Principle to transform this problem into two sub-problems, one for link selection and the other for relay assignment when there is enough replays. We prove that the proposed scheme can achieve an optimality gap of just 1 time slot at greatly reduced complexity. We also develop a heuristic scheme to handle the case when there is no enough relays. The proposed schemes are validated with simulations, where their superior performance is observed.

We investigate the problem of streaming multi-user videos over the downlink of a cognitive radio network (CRN), where each cognitive user (CU) can access one channel at a time. We first consider the case where each CU can sense one channel at a time slot at most. To make the problem tractable, we tackle the optimal spectrum sensing and access problems separately and develop matching-based optimal algorithms to the subproblems, which yield an overall suboptimal solution. We then consider the case where each CU can sense multiple channels. We show that under the assumption that all the spectrum sensors work on the same operating point, a two-step approach can derive the optimal spectrum sensing and access policies that maximize the quality of experience (QoE) of the streaming videos. The superior performance of the proposed approaches is validated with simulations and comparisons with benchmark schemes, where a performance gain from 25% to 30% is demonstrated.

We tackle the challenging problem of streaming multiuser videos over the downlink of a cellular cognitive radio network (CRN), where each cognitive user (CU) can sense and access multiple channels at a time. Spectrum sensing, channel assignment, and power allocation strategies are jointly optimized to maximize the quality of service (QoS) for the CUs. We show that the formulated mixed integer nonlinear programming (MINLP) problem can be decomposed into two subproblems: 1) SP1 for the optimal spectrum sensing strategy and 2) SP2 for the optimal channel assignment and power allocation, without sacrificing optimality. We show that SP1 can be optimally solved if there is no restriction on the sensing capability for each CU, and develop a column generation (CG)-based algorithm to solve SP2 iteratively in a distributed manner. We also develop a heuristic algorithm for spectrum sensing with greatly reduced requirement on CU hardware, while still achieving a highly competitive sensing performance. We analyze the proposed algorithms with respect to complexity and time efficiency, and derive a performance upper bound. The proposed algorithms are validated with simulations.

We study the problem of minimizing the scheduling time length to serve users' traffic demand by link scheduling in multi-hop mmWave wireless networks. We formulate a constrained Binary Integer Programming (BIP) problem incorporating a flexible interference model for directional transmissions and a Markov chain based blockage model. Since the problem is NP hard, we propose a heuristic algorithm with greatly reduced complexity, which first finds the optimal streaming path for each data flow and then maximizes the instant network throughput by optimizing the link scheduling at each time slot. The performance of the heuristic algorithm is validated with simulations.

In this paper we tackle the problem of minimum time length link scheduling in 60GHz wireless networks, under both traffic demand and Signal to Interference and Noise Ratio (SINR) constraints. A constrained Binary Integer Programming (BIP) problem is formulated by incorporating a flexible interference model for directional transmissions and a Markov chain based blockage model. We then propose two effective solution algorithms, including a Greedy Algorithm (GA) that finds the maximum instant throughput for each time slot, and a Column Generation based algorithm (CG) that iteratively improves the current link schedule. The performance of the proposed algorithms is validated with simulations.

We consider the problem of streaming multi-user videos over the downlink of a Cognitive Radio Network (CRN), where each Cognitive User (CU) can access one channel at a time. Motivated by the prior work that establishes a separation principle for the joint design of spectrum sensor, sensing, and access polices, we first model cooperative spectrum sensing as an Integer Programming problem (IP) and develop a Greedy Poly-matching scheme to solve it for the optimal sensing strategies. We then formulate the problem of CU Quality of Experience (QoE) maximization as a maximum weight matching problem and solve it with the Hungarian Method for optimal channel assignments. The proposed spectrum sensing and channel assignment algorithms are compared with benchmark schemes in simulations, and are found to outperform the benchmark schemes in terms of available channels discovered and CU QoE achieved.

We tackle the challenging problem of streaming multi-user videos over the downlink of a cellular Cognitive Radio Network (CRN), where each Cognitive User (CU) can sense and access multiple channels at a time. Spectrum sensing, channel assignment, and power allocation strategies are jointly optimized to maximize the Quality of Service (QoS) for the CUs. We show that the formulated Mixed Integer NonLinear Programming (MINLP) problem can be decomposed into two sub-problems without sacrificing optimality: SP1 for the optimal spectrum sensing strategy, and SP2 for optimal channel assignment and power allocation. We show that SP1 can be optimally solved and then develop a Column Generation (CG) based algorithm to solve SP2 iteratively in a distributed manner. We also develop a heuristic algorithm for spectrum sensing with greatly reduced requirement on CU hardware, but with a highly competitive sensing performance. We analyze the proposed algorithms with respect to complexity and derive a performance upper bound. The proposed algorithms are validated with extensive simulations.