|
Learning multiple metrics for ranking
Xiubo GENG, Xue-Qi CHENG
Front Comput Sci Chin. 2011, 5 (3): 259-267.
https://doi.org/10.1007/s11704-011-0152-5
Directly optimizing an information retrieval (IR) metric has become a hot topic in the field of learning to rank. Conventional wisdom believes that it is better to train for the loss function on which will be used for evaluation. But we often observe different results in reality. For example, directly optimizing averaged precision achieves higher performance than directly optimizing precision@3 when the ranking results are evaluated in terms of precision@3. This motivates us to combine multiple metrics in the process of optimizing IR metrics. For simplicity we study learning with two metrics. Since we usually conduct the learning process in a restricted hypothesis space, e.g., linear hypothesis space, it is usually difficult to maximize both metrics at the same time. To tackle this problem, we propose a relaxed approach in this paper. Specifically, we incorporate one metric within the constraint while maximizing the other one. By restricting the feasible hypothesis space, we can get a more robust ranking model. Empirical results on the benchmark data set LETOR show that the relaxed approach is superior to the direct linear combination approach, and also outperforms other baselines.
Figures and Tables |
References |
Related Articles |
Metrics
|
|
Traceback in wireless sensor networks with packet marking and logging
Jun XU, Xuehai ZHOU, Feng YANG
Front Comput Sci Chin. 2011, 5 (3): 308-315.
https://doi.org/10.1007/s11704-011-0361-y
In a hostile environment, sensor nodes may be compromised and then be used to launch various attacks. One severe attack is false data injection which is becoming a serious threat to wireless sensor networks. An attacker uses the compromised node to flood the network and exhaust network resources by injecting a large number of bogus packets. In this paper, we study how to locate the attack node using a framework of packet marking and packet logging. We propose a combined packet marking and logging scheme for traceback (CPMLT). In CPMLT, one packet can be marked by up to M nodes, each node marks a packet with certain probability. When one packet is marked by M nodes, the next marking node will log this packet. Through combining packet marking and logging, we can reconstruct the entire attack path to locate the attack node by collecting enough packets. In our simulation, CPMLT achieves fast traceback with little logging overhead.
References |
Related Articles |
Metrics
|
|
Reasonable routing in delay/disruption tolerant networks
Haizheng YU, Jianfeng MA, Hong BIAN
Front Comput Sci Chin. 2011, 5 (3): 327-334.
https://doi.org/10.1007/s11704-011-0139-2
Delay/disruption tolerant networking (DTN) is an approach to networking where intermittent connectivity exists: it is often afforded by a store and forward technique. Depending on the capability of intermediary nodes to carry and forward messages, messages can be eventually delivered to their destination by mobile nodes with an appropriate routing protocol. To have achieved a successful delivery, most DTN routing protocols use message duplication methods. Although messages are rapidly transferred to the destination, the redundancy in the number of message copies increases rapidly. This paper presents a new routing scheme based on a stochastic process for epidemic routing. Message redundancy is efficiently reduced and the number of message copies is controlled reasonably. During the contact process of nodes in the network, the number of message copies changes, and according to the variability in the number of copies, we construct a special Markov chain, birth and death process, on the number of message copies then calculate and obtain a stationary distribution of the birth and death process. Comparing the theoretical model with the simulation we have performed we see similar results. Our method improves on time-to-live (TTL) and anti-packet methods, in both redundancy and delivery success efficiency.
References |
Related Articles |
Metrics
|
|
Green challenges to system software in data centers
Yuzhong SUN, Yiqiang ZHAO, Ying SONG, Yajun YANG, Haifeng FANG, Hongyong ZANG, Yaqiong LI, Yunwei GAO
Front Comput Sci Chin. 2011, 5 (3): 353-368.
https://doi.org/10.1007/s11704-011-0369-3
With the increasing demand and the wide application of high performance commodity multi-core processors, both the quantity and scale of data centers grow dramatically and they bring heavy energy consumption. Researchers and engineers have applied much effort to reducing hardware energy consumption, but software is the true consumer of power and another key in making better use of energy. System software is critical to better energy utilization, because it is not only the manager of hardware but also the bridge and platform between applications and hardware. In this paper, we summarize some trends that can affect the efficiency of data centers. Meanwhile, we investigate the causes of software inefficiency. Based on these studies, major technical challenges and corresponding possible solutions to attain green system software in programmability, scalability, efficiency and software architecture are discussed. Finally, some of our research progress on trusted energy efficient system software is briefly introduced.
Figures and Tables |
References |
Related Articles |
Metrics
|
|
Static typing for a substructural lambda calculus
Baojian HUA
Front Comput Sci Chin. 2011, 5 (3): 369-380.
https://doi.org/10.1007/s11704-011-9106-1
Substructural type systems are designed from the insight inspired by the development of linear and substructural logics. Substructural type systems promise to control the usage of computational resources statically, thus detect more program errors at an early stage than traditional type systems do. In the past decade, substructural type systems have been deployed in the design of novel programming languages, such as Vault, etc. This paper presents a general typing theory for substructural type system. First, we define a universal semantic framework for substructural types by interpreting them as characteristic intervals composed of type qualifiers. Based on this framework, we present the design of a substructural calculus λSL with subtyping relations. After giving syntax, typing rules and operational semantics for λSL, we prove the type safety theorem. The new calculus λSL can guarantee many more safety invariants than traditional lambda calculus, which is demonstrated by showing that the λSL calculus can serve as an idealized type intermediate language, and defining a type- preserving translation from ordinary typed lambda calculus into λSL.
Figures and Tables |
References |
Related Articles |
Metrics
|
12 articles
|