教学课程资源库 更多>>

关键词： Computer science  

摘要： Sparse matrix computations, in the form of solvers for systems of linear equations, eigenvalue problem or matrix factorizations constitute the main kernel in problems from fields as diverse as computational fluid dynamics, quantum many body problems, machine learning and graph analytics. Iterative eigensolvers have been preferred over the regular method because the regular method not being feasible with industrial sized matrices. Although dense linear algebra libraries like BLAS, LAPACK, SCALAPACK are well established and some vendor optimized implementation like mkl from Intel or Cray Libsci exist, it is not the same case for sparse linear algebra which is lagging far behind. The main reason behind slow progress in the standardization of sparse linear algebra or library development is the different forms and properties depending on the application area. It is worsened for deep memory hierarchies of modern architectures due to low arithmetic intensities and memory bound computations. Minimization of data movement and fast access to the matrix are critical in this case. Since the current technology is driven by deep memory architectures where we get the increased capacity at the expense of increased latency and decreased bandwidth when we go further from the processors. The key to achieve high performance in sparse matrix computations in deep memory hierarchy is to minimize data movement across layers of the memory and overlap data movement with computations. My thesis work contributes towards addressing the algorithmic challenges and developing a computational infrastructure to achieve high performance in scientific applications for both shared memory and distributed memory architectures. For this purpose, I started working on optimizing a blocked eigensolver and optimized specific computational kernels which uses a new storage format. Using this optimization as a building block, we introduce a shared memory task parallel framework focusing on optimizing the entire

关键词： Chemical engineering   Polymer chemistry   Materials science   Computer science   Pharmaceutical sciences   Design   Neurosciences  

摘要： Data-driven design and study of new materials and (pre)clinical drugs have been increasing popularity with advancement of machine learning, data science, molecular simulation, and high-performance computation. To overcome the current experimental “trial-and-error” strategy for property evaluation and design of functional soft materials and pharmaceutical drugs, we developed a data-driven computational approaches for rational design and discovery antifouling materials and amyloid inhibitors. Antifouling materials and coatings have an increasing fundamental and practical applications. Specifically, we computationally studied antifouling mechanisms of (1) four different acrylamides (AMs) for their interfacial water behaviors and their interactions with a protein using all-atom, explicit solvent molecular dynamics MD simulations and (2) four poly(N-hydroxyalkyl acrylamide) (PAMs) brushes and (3) three zwitterionic brushes using a combination of molecular mechanics (MM), Monte Carlo (MC), and MD simulations. These works provide some structural insights into the design of new antifouling materials and surfaces. Beyond antifouling mechanism investigation, we developed a data-driven computational program for rational design and discovery of antifouling self-assembled monolayers (SAMs) and polymer brushes using machine learning algorithms, reveling molecular descriptors and functional groups important for antifouling properties of materials. These data-driven machine learning models can be used as an intelligent tool for determining, repurposing, and designing new superior antifouling materials and surfaces. Amyloid aggregation and formation are the common pathological characteristics of different neurodegenerative diseases involving Alzheimer disease (AD) and type II diabetes (T2D). However, currently developed amyloid inhibitors can’t achieve clinic success due to their poor biocompatibility low binding affinity and selectivity, and low permeability across blood-brain barr

关键词： Computer science   Transportation   Computer engineering   Automotive engineering  

摘要： Vehicular Ad Hoc Network (VANET) is a pervasive network where vehicles communicate with nearby vehicles and infrastructure nodes, such as Road-side unit (RSU). VANET is the subclass of Mobile Ad Hoc Network (MANET) in which nodes move randomly and are connected wirelessly. Information sharing among vehicles is an essential component of an intelligent traffic system (ITS), but security and privacy concerns must be taken into consideration. Security of the network can be improved by granting access only to authenticated vehicles. This research proposes an RSU based approach to authenticate vehicles and notify vehicles about unauthorized messages/vehicles. It helps in preventing other vehicles in the network from being influenced by the malicious vehicle. In this approach, Blockchain has been used to securely maintain the identity of all vehicles in the network. The use of this RSU based approach helps to reduce the computational overhead on the On-board unit (OBU) of individual vehicles and reduces the processing delay.

关键词： Sensitivity analysis   Artificial intelligence   Adaptation   Algorithms   Ablation   Distance learning   Computer science  

摘要： Artificial intelligence questions our understanding of intelligent behaviour and our own intelligence. Recent advances rely on machine learning, where intelligence arises through statistical learning. Often, machine learning assumes agents approach tasks with no prior knowledge. This stands in stark contrast to how humans approach new problems and is unlikely to yield human-level intelligence that can learn to solve new, unseen problems as they arise. To this end, we need a learning paradigm at a higher level of abstraction. One alternative is agents that learn to learn. Within this framework, agents learn not to solve a given set of tasks, but how to solve them. Such agents can generalise prior experiences into abstract concepts for learning and problem solving. Within the context of neural networks, current methods are limited in their ability to generalise and scale in terms of task variation and complexity. This thesis makes four contributions that tackle these challenges. A novel method is proposed that learns to dynamically adapt an agent's parameters to increase its expressive capacity and ability to generalise. Further, a framework is proposed that grounds meta-learning in differential geometry by learning shortest solution paths (geodesics) across tasks. Building on these insights, a novel method for meta-learning is proposed that is simple, scalable, and effective. It is the first gradient-based meta-learner that can be directly applied to any form of learningâ--including supervised, unsupervised, reinforcement, continual, and online learningâ€--opening up for meta-learning at the scale and at the level of complexity required for sophisticated artificial intelligence. Finally, while the above methods rely on a predefined task distribution, an artificial intelligence should be able to define its own tasks as needed. To this end, a novel system for exploration in reinforcement learning is proposed that creates intrinsic tasks. These drive an agent to explore

关键词： Bioinformatics   Computer science   Biology  

摘要： During my Ph.D., I developed several computational approaches to advance precision medicine for cancer prevention and treatment. My thesis presents three such approaches addressing these emerging challenges by analyzing large-scale cancer omics data from both pre-clinical models and patients datasets. In the first project, we studied the cancer risk associated with CRISPR-based therapies. Therapeutics based on CRISPR technologies (for which the chemistry Nobel prize was awarded in 2020) are poised to become widely applicable for treating a variety of human genetic diseases. However, preceding our work, two experimental studies have reported that genome editing by CRISPR–Cas9 can induce a DNA damage response mediated by p53 in primary cells hampering their growth. This could lead to an undesired selection of cells with pre-existing p53 mutations. Motivated by these findings, we conducted the first comprehensive computational and experimental investigation of the risk of CRISPR-induced selection of cancer gene mutants across many different cell types and lineages. I further studied whether this selection is dependent on the Cas9/sgRNA-delivery method and/or the gene being targeted. Importantly, we asked whether other cancer driver mutations may also be selected during CRISPR-Cas9 gene editing and identified that pre-existing KRAS mutants may also be selected for during CRISPR-Cas9 editing. In summary, we established that the risk of selection for pre-existing p53 or KRAS mutations is non-negligible, thus calling for careful monitoring of patients undergoing CRISPR-Cas9-based editing for clinical therapeutics for pre-existing p53 and KRAS mutations. In the second project, we aimed to delineate some of the molecular mechanisms that may underlie the observed differences in cancer incidences across cancer patients of different ancestries, focusing mainly on lung cancer. We found that lung tumors from African American (AA) patients exhibit higher genomic instability, homol

关键词： Kinematics   Control algorithms   Collaboration   Traffic   Cooperation   Planning   Optimization   Controllers   Robots   Connectivity   Vehicles   Computer science   Mechanics   Physics   Robotics   Transportation  

摘要： Platooning of connected and automated vehicles (CAVs) has received extensive interest due to its potential to improve road traffic. In this context, multi-lane platoons have appeared as a generalized version of the classical train-like platoon structure used in the early platoon implementations from the 80s. Multi-lane platoons add structural flexibility to perform complex inter-vehicle maneuvers such as lane changes, coordination during bottlenecks, among others. Nonetheless, these advantages are at the price of more intricate internal dynamics. This work addresses motion planning in tight multi-lane platoons using distributed techniques that make use of a graph theoretical formulation of platoon formations, for reference generation, and distributed non-linear model predictive control, for vehicle command. The effectiveness of the proposed approaches is tested by simulating a variety of scenarios with progressive complexity, where it is shown that graph-based distributed approaches inherently promote cooperation within vehicles. Moreover, the second proposed approach, denominated incremental approach, is robust to sensor biases

关键词： Computer science   Environmental science  

关键词： 多样性   理论计算机科学   算法   体外化   人类世  

摘要： 斯蒂格勒认为,我们这个数字化时代正是充满技术所释放的毒性的人类世时代,而且正处于其毒性最大化的时期。他的这个论断正在为2020年初突如其来的、影响全球各个国家的新冠疫情所强化。而当今人类世之所以出现如此的局面,是因为人类自启蒙运动以来过于相信科学技术所致。在斯蒂格勒看来,人类的进化是体外进化的模式,在这个进化过程中,人类的生存会不断地更加依赖外在于其躯体的技术和技术物体,也即斯蒂格勒所说的第三滞留。工业革命推动了科学技术的发展,而在此过程中出现的以人工智能的诸种形式为伪装的数字技术,以异常强大的力量将人类躯体内器官的感知、知性、理性等心智功能不断地外化于人类躯体之外的所谓高新技术物之中。这些技术物如今已到处雷同泛滥,它们是提示记忆的第三滞留的最新形式。于是,雷同的提示记忆的方式导致了雷同的心智思维方式,而雷同的心智思维方式导致了系统性的愚蠢,导致了人类熵在全球范围内的急速增长。在此意义上,斯蒂格勒才说,人类世已经达到其毒性最大化的时期。当然,斯蒂格勒并非是完全贬斥技术的人,因为在他的理论中,技术既是毒药也是解药。当今数字化的科学技术导致了人类世毒性的最大化,但要化解这种毒性却依然需要依赖于这些数字科技。这也就是斯蒂格勒为什么在这里要重新反思理论计算机科学之基础的原因。只有主动地反思科学技术才能构建心智的多样性,才能认识到科学技术并不是唯一的技术形式,进而才能保护并发展技术的多样性。斯蒂格勒似乎将这场疫情看作是能够促使人类反思当下人类世状态的契机。可是,他本人却又对当下人类世的诸种状况非常地失望,他在这个大流行病的2020年结束了自己的生命。哲学家并非一定是相信自己哲学理论的人,也并非一定是勇于践行自身哲学理论的人。苏格拉底相信并践行了自己的哲学理论。那么,于斯蒂格勒而言,其为是耶?其为非耶?

关键词： Computer engineering   Computer science   Electrical engineering   Artificial intelligence  

摘要： Artificial intelligence (AI) has experienced a tremendous surge in recent years, resulting in high demand for a wide array of implementations of algorithms in the field. With the rise of Internet-of-Things devices, the need for artificial intelligence algorithms implemented in hardware with tight design restrictions has become even more prevalent. In terms of low power and area, ASIC implementations have the best case. However, these implementations suffer from high non-recurring engineering costs, long time-to-market, and a complete lack of flexibility, which significantly hurts their appeal in an environment where time-to-market is so critical. The time-to-market gap can be shortened through the use of reconfigurable solutions, such as FPGAs, but these come with high cost per unit and significant power and area deficiencies over their ASIC counterparts. To bridge these gaps, this dissertation work develops two methodologies to improve the usability of ASIC implementations of neural networks in these applications. The first method demonstrates a method for substantial reductions in design time for asynchronous implementations of a set of AI algorithms known as Recurrent Neural Networks (RNN) by analyzing the possible architectures and implementing a library of generic or easily altered components that can be used to quickly implement a chosen RNN architecture. A tapeout of this method was completed using as few as 112 hours of labor by the designer from RNN selection to a DRC/LVS clean chip layout ready for fabrication. The second method develops a flow to implement a set of RNNs in a single reconfigurable ASIC, offering a middle ground between fully reconfigurable solutions and completely application-specific implementations. This reconfigurable design is capable of representing thousands of possible RNN configurations in a single IC. A tapeout of this design was also completed, with both tapeouts using the TSMC 65nm bulk CMOS process.

关键词： Aerospace engineering   Artificial intelligence   Computer science   Engineering   Operations research  

摘要： Autonomous agents have the potential to do tasks that would otherwise be too repetitive, difficult, or dangerous for humans. Solving many of these problems requires reasoning over sequences of decisions in order to reach a goal. Autonomous driving, inventory management, and medical diagnosis and treatment are all examples of important real-world sequential decision problems. Approximate solution methods such as reinforcement learning and Monte Carlo planning have achieved superhuman performance in some domains. In these methods, agents learn good actions to take in response to inputs. Problems with many widely varying inputs or possible actions remain challenging to efficiently solve without extensive problem-specific engineering. One of the key challenges in solving sequential decision problems is efficiently exploring the many different paths an agent may take. For most problems, it is infeasible to test every possible path. Many existing approaches explore paths using simple random sampling. Problems in which many different actions may be taken at each step often require more efficient exploration to be solved. Large, unstructured input spaces can also challenge conventional learning approaches. Agents must learn to recognize inputs that are functionally similar while simultaneously learning an effective decision strategy. As a result of these challenges, learning agents are often limited to solving tasks in virtual domains where very large amounts of trials can be conducted relatively safely and cheaply. When problems are solved using black-box models such as neural networks, the resulting decision making policy is impossible for a human to meaningfully interpret. This can also limit the use of learning agents to low-regret tasks such as image classification or video game playing. The work in this thesis addresses the challenges of learning in large-space sequential decision problems. The thesis first considers methods to improve scaling of deep reinforcement le