Improving sentiment analysis via sentence type classification using BiLSTM-CRF and CNN

Different types of sentences express sentiment in very different ways. Traditional sentence-level sentiment classification research focuses on one-technique-fits-all solution or only centers on one special type of sentences. In this paper, we propose a divide-and-conquer approach which first classifies sentences into different types, then performs sentiment analysis separately on sentences from each type. Specifically, we find that sentences tend to be more complex if they contain more sentiment targets. Thus, we propose to first apply a neural network based sequence model to classify opinionated sentences into three types according to the number of targets appeared in a sentence. Each group of sentences is then fed into a one-dimensional convolutional neural network separately for sentiment classification. Our approach has been evaluated on four sentiment classification datasets and compared with a wide range of baselines. Experimental results show that: (1) sentence type classification can improve the performance of sentence-level sentiment analysis; (2) the proposed approach achieves state-of-the-art results on several benchmarking datasets.

Statistical mechanics of low-density parity-check codes

We review recent theoretical progress on the statistical mechanics of error correcting codes, focusing on low-density parity-check (LDPC) codes in general, and on Gallager and MacKay-Neal codes in particular. By exploiting the relation between LDPC codes and Ising spin systems with multispin interactions, one can carry out a statistical mechanics based analysis that determines the practical and theoretical limitations of various code constructions, corresponding to dynamical and thermodynamical transitions, respectively, as well as the behaviour of error-exponents averaged over the corresponding code ensemble as a function of channel noise. We also contrast the results obtained using methods of statistical mechanics with those derived in the information theory literature, and show how these methods can be generalized to include other channel types and related communication problems.

Suppression and summation in contrast gain control for human vision

Over the last ten years our understanding of early spatial vision has improved enormously. The long-standing model of probability summation amongst multiple independent mechanisms with static output nonlinearities responsible for masking is obsolete. It has been replaced by a much more complex network of additive, suppressive, and facilitatory interactions and nonlinearities across eyes, area, spatial frequency, and orientation that extend well beyond the classical recep-tive field (CRF). A review of a substantial body of psychophysical work performed by ourselves (20 papers), and others, leads us to the following tentative account of the processing path for signal contrast. The first suppression stage is monocular, isotropic, non-adaptable, accelerates with RMS contrast, most potent for low spatial and high temporal frequencies, and extends slightly beyond the CRF. Second and third stages of suppression are difficult to disentangle but are possibly pre- and post-binocular summation, and involve components that are scale invariant, isotropic, anisotropic, chromatic, achromatic, adaptable, interocular, substantially larger than the CRF, and saturated by contrast. The monocular excitatory pathways begin with half-wave rectification, followed by a preliminary stage of half-binocular summation, a square-law transducer, full binocular summation, pooling over phase, cross-mechanism facilitatory interactions, additive noise, linear summation over area, and a slightly uncertain decision-maker. The purpose of each of these interactions is far from clear, but the system benefits from area and binocular summation of weak contrast signals as well as area and ocularity invariances above threshold (a herd of zebras doesn't change its contrast when it increases in number or when you close one eye). One of many remaining challenges is to determine the stage or stages of spatial tuning in the excitatory pathway.

The co-ordination chemistry of tellurium and related selenium ligands

2-(2-pyridyl)phenyl(p-ethoxyphenyl)tellurium(II), (RR1Te) reacts with HgC12 at room temperature to give white HgCl2.RR1Te. On setting aside, or on warming the reaction mixture a yellow material, [R1HgCl.(RTeCl)2] is formed. Multinuclear NMR(125Te, 199Hg, 1H) and mass spectroscopy confirm the formulation, and confirm the ease of transfer of the p-ethoxyphenyl group (R1) between the metal centres. The crystal structure of the yellow material consists of two discrete RTeCl molecules together with a R1HgCl molecule. There is no dative bond formation between these species, hence the preferred description of the formation of an inclusion complex. The reaction of RR1Te with Copper(I) chloride in the cold gives an air sensitive yellow product Cu3Cl3(RR1Te)2(0.5CH3CN); under reflux in air changes to the green Cu2Cl(RR1Te)(0.5 EtOH). By contrast, the reaction of RR1Te with acetonitrile solution of Copper(II) salts under mild conditions affords the white materials CuCl(RR1Te) and CuBr(RR1Te)H2O. RR1Te reacts with PdCl2 and PtCl2 to give materials albeit not well defined, can be seen as intermediates to the synthesis of inorganic phase of the type M3XTe2XCl2X. Paramagnetism is associated with some of the palladium and platinum products. The 195Pt NMR measurement in DMSO establishes the presence of six platinum species, which are assigned to Pt(IV), Pt(III) or Pt(II). The reactions show that in the presence of PdCl2 or PtCl2 both R and R1 are very labile. The reaction of RHgCl(R= 2-(2-pyridyl)phenyl) with SeX4(X= Cl, Br) gives compounds which suggest that both Trans-metallation and redox processes are involved. By varying reaction conditions materials which appear to be intermediates in the trans-metallation process are isolated. Potentially bidentate tellurium ligands having molecular formula RTe(CH2)nTeR,Ln, (R= Ph,(t-Bu). C6H4, n = 5,10) are prepared. Palladium and Platinum complexes containing these ligands are prepared. Also complex Ph3SnC1L(L = p-EtO.C6H4) is prepared.

Growth curve of the lichen Rhizocarpon geographicum

Data on the growth curve of the lichen Rhizocarpon geographicum were obtained by measuring the radial growth rates (mm per 1.5 years) of 39 thalli from 2 to 65 mm in diameter growing in the same environment. An Aplin and Hill plot (r2 – r1 against ln r2 – ln r1) of the data and regression analyses suggested an initial phase of growth (up to a diameter of about 7 mm) in which the relative growth rate increased rapidly. This was followed by a phase in which the relative growth rate fell but the radial growth rate continued to rise (7 to 20 mm in diameter). Radial growth was then relatively constant until about 45 mm diameter and then declined. The Aplin and Hill model did not fit the data as a whole but may apply for a transient period in thalli between about 7 and 16 mm in diameter. The curve shows some similarities to that suggested by lichenometric studies but differs in showing a less steep decline in growth rate after the ‘great’ period.

Solvent induced NMR chemical shifts that arise from molecular encounters

Recently Homer and Percival have postulated that intermolecular van der Waals dispersion forces can be characterized by three mechanisms. The first arises via the mean square reaction field < R1; 2> due to the transient dipole of a particular solute molecule that is considered situated in a cavity surrounded by solvent molecules; this was characterized by an extended Onsager approach. The second stems from the extra cavity mean square reaction field < R2; 2> of the near neighbour solvent molecules. The third originates from square field electric fields E2BI due to a newly characterized effect in which solute atoms are `buffeted' by the peripheral atoms of adjacent solvent molecules. The present work concerns more detailed studies of the buffeting screening, which is governed by sterically controlled parameter (2T - T)2, where and are geometric structural parameters. The original approach is used to characterise the buffeting shifts induced by large solvent molecules and the approach is found to be inadequate. Consequently, improved methods of calculating and are reported. Using the improved approach it is shown that buffeting is dependent on the nature of the solvent as well as the nature of the solute molecule. Detailed investigation of the buffeting component of the van der Waals chemical shifts of selected solutes in a range of solvents containing either H or Cl as peripheral atoms have enabled the determination of a theoretical acceptable value for the classical screening coefficient B for protons. 1H and 13C resonance studies of tetraethylmethane and 1H, 13C and 29Si resonance studies of TMS have been used to support the original contention that three (< R1; 2> , < R2; 2> and E2BI) components of intermolecular van der Waals dispersion fields are required to characterise vdW chemical shifts.

Insulin release from cultured B-cells

Established RlNm5F and lN111 R1 and newly available HlT-T15 and UMR 407/3 B-cell lines have been successfully maintained in vitro. With the exclusion of UMR 407/3 cells, all lines were continuously propagable. Doubling times and plating efficiencies for HlT-T15, RlNm5F, lN111 R1 and UMR 407/3 cells were 20 hours and 85%, 31 hours and 76%, 24 hours and 80% and 38 hours and 94% respectively. All the cell lines were anchorage dependent, but only UMR 407/3 cells grew to confluence. Only HlT-T15 and UMR 407/3 cells produced a true insulin response to glucose but glucose markedly increased the rate of D-[U14C]glucose oxidation by all the cell lines. Glucose induced insulin release from HlT-T15 cells was biphasic with an exaggerated first phase. Insulin release from HlT-T15, RlNm5F and IN111 R1 cells was stimulated by amino acids and sulphonylureas. Glucagon stimulated insulin release from HlT-T15 and RlNm5F cells while somatostatin and pancreatic polypeptide inhibited release. These observations suggest that net insulin release from the whole islet may be the result of significant paracrine interaction. HlT-T15 and RlNm5F cell insulin release was stimulated by forskolin and inhibited by imidazole. Ca2+ channel blockade and calmodulin inhibition suppressed insulin release from HlT-T15, RlNm5F and IN111 R1 cells. In addition phorbol esters stimulated insulin release from RlNm5F cells. These data implicate cAMP, Ca2+ and protein kinase-C in the regulation of insulin release from cultured B-cells. Acetylcholine increased insulin release from HlT-T15 and RlNm5F cells. Inhibition of the response by atropine confirmed the involvement of muscarinic receptors. HlT-T15 cell insulin release was also inhibited by adrenaline. These observations suggest a possible role for the autonomic nervous system in the modulation of insulin release. Preliminary studies with a human insulinoma maintained in monolayer culture have demonstrated a limited life span of some seven weeks, a continuous low level of insulin release but no insulin response to glucose challenge.

Low thermal sensitivity grating devices based on ex-45° tilting structure capable of forward-propagating cladding modes coupling

The authors describe a detailed investigation on tilted fiber Bragg grating (TFBG) structures with tilted angles exceeding 45°. In contrast to the backward mode coupling mechanism of Bragg gratings with normal and small tilting structures, the ex-45° TFBGs facilitate the light coupling to the forward-propagating cladding modes. The authors have also theoretically and experimentally examined the mode coupling transition of TFBGs with small, medium, and large tilt angles. In particular, experiments are conducted to investigate the spectra and far-field distribution, as well as temperature, strain, and refractive-index sensitivities of ex-45° devices. It has been revealed that these ex-45° gratings exhibit ultralow thermal sensitivity. As in-fiber devices, they may be superior to conventional Bragg and long-period gratings when the low thermal cross sensitivity is required.

Accuracy of Goldmann, ocular response analyser, Pascal and TonoPen XL tonometry in keratoconic and normal eyes

Aim: The aim of this study was to evaluate the practicality and accuracy of tonometers used in routine clinical practice for established keratoconus (KC). Methods: This was a prospective study of 118 normal and 76 keratoconic eyes where intraocular pressure (IOP) was measured in random order using the Goldman applanation tonometer (GAT), Pascal dynamic contour tonometer (DCT), Reichert ocular response analyser (ORA) and TonoPen XL tonometer. Corneal hysteresis (CH) and corneal resistance factor (CRF), as calculated by the ORA, were recorded. Central corneal thickness (CCT) was measured using an ultrasound pachymeter. Results: The difference in IOP values between instruments was highly significant in both study groups (p<0.001). All other IOP measures were significantly higher than those for GAT, except for the Goldmann-correlated IOP (average of the two applanation pressure points) (IOPg) as measured by ORA in the control group and the CH-corrected IOP (corneal-compensated IOP value) (IOPcc) measures in the KC group. CCT, CH and CRF were significantly less in the KC group (p<0.001). Apart from the DCT, all techniques tended to measure IOP higher in eyes with thicker corneas. Conclusion: The DCT and the ORA are currently the most appropriate tonometers to use in KC for the measurement of IOPcc. Corneal factors such as CH and CRT may be of more importance than CCT in causing inaccuracies in applanation tonometry techniques.

Receptor Activity Modifying Proteins (RAMPs) interact with the VPAC2 Receptor and CRF1 receptors and modulate their function

Background and Purpose Although it is established that the receptor activity modifying proteins (RAMPs) can interact with a number of GPCRs, little is known about the consequences of these interactions. Here the interaction of RAMPs with the glucagon-like peptide 1 receptor (GLP-1 receptor), the human vasoactive intestinal polypeptide/pituitary AC-Activating peptide 2 receptor (VPAC) and the type 1 corticotrophin releasing factor receptor (CRF) has been examined. Experimental Approach GPCRs were co-transfected with RAMPs in HEK 293S and CHO-K1 cells. Cell surface expression of RAMPs and GPCRs was examined by elisa. Where there was evidence for interactions, agonist-stimulated cAMP production, Ca mobilization and GTPγS binding to G, G, G and G were examined. The ability of CRF to stimulate adrenal corticotrophic hormone release in Ramp2 mice was assessed. Key Results The GLP-1 receptor failed to enhance the cell surface expression of any RAMP. VPAC enhanced the cell surface expression of all three RAMPs. CRF enhanced the cell surface expression of RAMP2; the cell surface expression of CRF was also increased. There was no effect on agonist-stimulated cAMP production. However, there was enhanced G-protein coupling in a receptor and agonist-dependent manner. The CRF: RAMP2 complex resulted in enhanced elevation of intracellular calcium to CRF and urocortin 1 but not sauvagine. In Ramp2 mice, there was a loss of responsiveness to CRF. Conclusions and Implications The VPAC and CRF receptors interact with RAMPs. This modulates G-protein coupling in an agonist-specific manner. For CRF, coupling to RAMP2 may be of physiological significance. © 2012 The Authors. British Journal of Pharmacology © 2012 The British Pharmacological Society.

The relationship between measurement method and corneal structure on apparent intraocular pressure in glaucoma and ocular hypertension

Purpose: To analyse the relationship between measured intraocular pressure (IOP) and central corneal thickness (CCT), corneal hysteresis (CH) and corneal resistance factor (CRF) in ocular hypertension (OHT), primary open-angle (POAG) and normal tension glaucoma (NTG) eyes using multiple tonometry devices. Methods: Right eyes of patients diagnosed with OHT (n=47), normal tension glaucoma (n=17) and POAG (n=50) were assessed, IOP was measured in random order with four devices: Goldmann applanation tonometry (GAT); Pascal(R) dynamic contour tonometer (DCT); Reichert(R) ocular response analyser (ORA); and Tono-Pen(R) XL. CCT was then measured using a hand-held ultrasonic pachymeter. CH and CRF were derived from the air pressure to corneal reflectance relationship of the ORA data. Results: Compared to the GAT, the Tonopen and ORA Goldmann equivalent (IOPg) and corneal compensated (IOPcc) measured higher IOP readings (F=19.351, p<0.001), particularly in NTG (F=12.604, p<0.001). DCT was closest to Goldmann IOP and had the lowest variance. CCT was significantly different (F=8.305, p<0.001) between the 3 conditions as was CH (F=6.854, p=0.002) and CRF (F=19.653, p<0.001). IOPcc measures were not affected by CCT. The DCT was generally not affected by corneal biomechanical factors. Conclusion: This study suggests that as the true pressure of the eye cannot be determined non-invasively, measurements from any tonometer should be interpreted with care, particularly when alterations in the corneal tissue are suspected.

Grading and disease management in national screening for diabetic retinopathy in England and Wales

Aims - A National Screening Programme for diabetic eye disease in the UK is in development. We propose a grading and early disease management protocol to detect sight-threatening diabetic retinopathy and any retinopathy, which will allow precise quality assurance at all steps while minimizing false-positive referral to the hospital eye service. Methods - Expert panel structured discussions between 2000 and 2002 with review of existing evidence and grading classifications. Proposals - Principles of the protocol include: separate grading of retinopathy and maculopathy, minimum number of steps, compatible with central monitoring, expandable for established more complex systems and for research, no lesion counting, no ‘questionable’ lesions, attempt to detect focal exudative, diffuse and ischaemic maculopathy and fast track referral from primary or secondary graders. Sight-threatening diabetic retinopathy is defined as: preproliferative retinopathy or worse, sight-threatening maculopathy and/or the presence of photocoagulation. In the centrally reported minimum data set retinopathy is graded into four levels: none (R0), background (R1), preproliferative (R2), proliferative (R3). Maculopathy and photocoagulation are graded as absent (M0, P0) or present (M1, P1). Discussion - The protocol developed by the Diabetic Retinopathy Grading and Disease Management Working Party represents a new consensus upon which national guidelines can be based leading to the introduction of quality-assured screening for people with diabetes.

Comparative quantitative study of ‘signature’ pathological lesions in the hippocampus and adjacent gyri of 12 neurodegenerative disorders

The hippocampus (HC) and adjacent gyri are implicated in dementia in several neurodegenerative disorders. To compare HC pathology among disorders, densities of ‘signature’ pathological lesions were measured at a standard location in eight brain regions of 12 disorders. Principal components analysis of the data suggested that the disorders could be divided into three groups: (1) Alzheimer’s disease (AD), Down’s syndrome (DS), sporadic Creutzfeldt–Jakob disease, and variant Creutzfeldt–Jakob disease in which either β-amyloid (Aβ) or prion protein deposits were distributed in all sectors of the HC and adjacent gyri, with high densities being recorded in the parahippocampal gyrus and subiculum; (2) Pick’s disease, sporadic frontotemporal lobar degeneration with TDP-43 immunoreactive inclusions, and neuronal intermediate filament inclusion disease in which relatively high densities of neuronal cytoplasmic inclusions were present in the dentate gyrus (DG) granule cells; and (3) Parkinson’s disease dementia, dementia with Lewy bodies, progressive supranuclear palsy, corticobasal degeneration, and multiple system atrophy in which densities of signature lesions were relatively low. Variation in density of signature lesions in DG granule cells and CA1 were the most important sources of neuropathological variation among disorders. Hence, HC and adjacent gyri are differentially affected in dementia reflecting either variation in vulnerability of hippocampal neurons to specific molecular pathologies or in the spread of pathological proteins to the HC. Information regarding the distribution of pathology could ultimately help to explain variations in different cognitive domains, such as memory, observed in various disorders.

Chain routing:a new routing framework for the Internet based on complete orders

A new framework to perform routing at the autonomous system (AS) level is proposed here. This mechanism, called chain routing framework (CRF), uses complete orders as its main topological unit. Since complete orders are acyclic digraphs that possess a known topology, it is possible to use these acyclic structures to route consistently packets between a group of ASs. The adoption of complete orders also allows easy identification and avoidance of persistent route oscillations, eliminates the possibility of developing transient loops in paths and provides a structure that facilitates the implementation of traffic engineering. Moreover, by combining CRF with other mechanisms that implement complete orders in time, the authors propose that it is possible to design a new routing protocol, which can be more reliable and stable than the border gateway protocol. © 2011 The Institution of Engineering and Technology.

Natural history of retinopathy in children and young people with type 1 diabetes

Purpose: To describe the prevalence and natural history of retinopathy in a cohort of children and young people with type 1 diabetes attending a tertiary hospital diabetes clinic. Methods: We analysed retinopathy screening data from 2008 to 2010 on all eligible children using the 'Twinkle' diabetes database and the regional retinal screening database. Results: A total of 88% (149/169) of eligible children were screened in 2008, median age 14 years, 52% male. The prevalence of retinopathy was 19.5% (30/149). All children had background retinopathy grade R1. There was significant difference in median (range) duration of diabetes, 7.7 years (0.6–13.7) vs 5 years (0.2–12.5) (P<0.001) and median (range) HbA1C, 9.1% (7.2–14) vs 8.6% (5.6–13.1) (P=0.02), between the groups with and without retinopathy. At 2- years follow-up, 12/30 (40%) had unchanged retinopathy grade R1, 10/30 (33.3%) showed resolution of changes (R0), 1/30 progressed to maculopathy, and 7/30 had no follow-up data. Median (range) HbA1C in 2008 and 2010 for the groups with stable vs resolved changes was similar, 9.1% (7.2–14.0) and 9.2% (7–14.0) vs 9.5% (7.8–14.0) and 9.2% (8.7–14.0). Of the 119 without retinopathy in 2008, 27 (22.5%) had developed retinopathy within 2 years, including 1 with pre-proliferative retinopathy and 1 with maculopathy. There was no significant difference in HbA1c between those who progressed to retinopathy (8.7% (7.1–13.1)) (8.7% (7.1–13.1)), and those who did not (8.6% (6.3–12.2)). Conclusions: Prevalence of background retinopathy in our cohort was comparable to the previously published reports, with higher HbA1c and longer duration of diabetes being significant risk factors. On short-term follow-up, Grade 1 retinopathy is likely to resolve in a third of patients and remain unchanged in just over a third.