The paper is concerned with applying learning to rank to document retrieval. Ranking SVM is a typical method of learning to rank. We point out that there are two factors one must consider when applying Ranking SVM, in general a "learning to rank" method, to document retrieval. First, correctly ranking documents on the top of the result list is crucial for an Information Retrieval system. One must conduct training in a way that such ranked results are accurate. Second, the number of relevant documents can vary from query to query. One must avoid training a model biased toward queries with a large number of relevant documents. Previously, when existing methods that include Ranking SVM were applied to document retrieval, none of the two factors was taken into consideration. We show it is possible to make modifications in conventional Ranking SVM, so it can be better used for document retrieval. Specifically, we modify the "Hinge Loss" function in Ranking SVM to deal with the problems described above. We employ two methods to conduct optimization on the loss function: gradient descent and quadratic programming. Experimental results show that our method, referred to as Ranking SVM for IR, can outperform the conventional Ranking SVM and other existing methods for document retrieval on two datasets.

In emerging optoelectronic applications, such as water photolysis, exciton fission and novel photovoltaics involving low-dimensional nanomaterials, hot-carrier relaxation and extraction mechanisms play an indispensable and intriguing role in their photo-electron conversion processes. Two-dimensional transition metal dichalcogenides have attracted much attention in above fields recently; however, insight into the relaxation mechanism of hot electron-hole pairs in the band nesting region denoted as C-excitons, remains elusive. Using MoS2 monolayers as a model two-dimensional transition metal dichalcogenide system, here we report a slower hot-carrier cooling for C-excitons, in comparison with band-edge excitons. We deduce that this effect arises from the favourable band alignment and transient excited-state Coulomb environment, rather than solely on quantum confinement in two-dimension systems. We identify the screening-sensitive bandgap renormalization for MoS2 monolayer/graphene heterostructures, and confirm the initial hot-carrier extraction for the C-exciton state with an unprecedented efficiency of 80%, accompanied by a twofold reduction in the exciton binding energy.

Graphene is regarded as a potential surface-enhanced Raman spectroscopy (SERS) substrate. However, the application of graphene quantum dots (GQDs) has had limited success due to material quality. Here, we develop a quasi-equilibrium plasma-enhanced chemical vapor deposition method to produce high-quality ultra-clean GQDs with sizes down to 2 nm directly on SiO2/Si, which are used as SERS substrates. The enhancement factor, which depends on the GQD size, is higher than conventional graphene sheets with sensitivity down to 1 × 10−9 mol L−1 rhodamine. This is attributed to the high-quality GQDs with atomically clean surfaces and large number of edges, as well as the enhanced charge transfer between molecules and GQDs with appropriate diameters due to the existence of Van Hove singularities in the electronic density of states. This work demonstrates a sensitive SERS substrate, and is valuable for applications of GQDs in graphene-based photonics and optoelectronics.