Learning Effective and Interpretable Semantic Models using Non-Negative Sparse Embedding

In this paper, we introduce an application of matrix factorization to produce corpus-derived, distributional
models of semantics that demonstrate cognitive plausibility. We find that word representations
learned by Non-Negative Sparse Embedding (NNSE), a variant of matrix factorization, are sparse,
effective, and highly interpretable. To the best of our knowledge, this is the first approach which
yields semantic representation of words satisfying these three desirable properties. Though extensive
experimental evaluations on multiple real-world tasks and datasets, we demonstrate the superiority
of semantic models learned by NNSE over other state-of-the-art baselines.

Commissioning and Evaluation of a Fiber-Optic Sensor System for Bridge Monitoring

This paper describes the design, commissioning, and evaluation of a ?ber-optic strain sensor system for the structural health monitoring of a prestressed concrete posttensioned box girder railway bridge in Mumbai, India, which shows a number of well-documented structural problems. Preliminary laboratory trials to design the most appropriate sensor system that could be readily transported and used on site are described, followed by a description of load tests on the actual bridge undertaken in collaboration with Indian Railways and using locomotives of known weight. Results from the load tests using the optical system are compared with similar results obtained using electrical resistance strain gages. Conclusions are summarized concerning the integrity of the structure and for the future use of the sensor system for monitoring bridges of this type. Crack width measurements obtained during the load tests are also described.

Chloride induced corrosion of steel bars in alkali activated slag concretes

The studies on chloride induced corrosion of steel bars in alkali activated slag (AAS) concretes are scarcely reported in the past. In order to make this issue clearer and compare the corrosion performance of AAS with Portland cement (PC) counterpart, an investigation was carried out and the results are reported in this paper. Corrosion properties were assessed with the help of rate of corrosion, electrical resistivity and pore solution chemistry. It was found that: (i) steel corrosion resistance of the AAS concretes was comparable or in some cases even worse than that of Portland cement (PC) concrete under intermittent chloride ponding regime; (ii) the corrosion behaviour of the AAS concretes was significantly influenced by ionic exchange, carbonation and sulphide concentration; (iii) the increase of alkali concentration of the activator generally reduced chloride resulting corrosion, and a value of 1.5 was found to be an optimum modulus for the activator for improving the corrosion resistance.