Novel Machine Learning HIV Intervention for Sexual and Gender Minority Young People Who Have Sex With Men (uTECH): Protocol for a Randomized Comparison Trial.

BACKGROUND: Sexual and gender minority (SGM) young people are disproportionately affected by HIV in the United States, and substance use is a major driver of new infections. People who use web-based venues to meet sex partners are more likely to report substance use, sexual risk behaviors, and sexually transmitted infections. To our knowledge, no machine learning (ML) interventions have been developed that use web-based and digital technologies to inform and personalize HIV and substance use prevention efforts for SGM young people. OBJECTIVE: This study aims to test the acceptability, appropriateness, and feasibility of the uTECH intervention, a SMS text messaging intervention using an ML algorithm to promote HIV prevention and substance use harm reduction among SGM people aged 18 to 29 years who have sex with men. This intervention will be compared to the Young Men's Health Project (YMHP) alone, an existing Centers for Disease Control and Prevention best evidence intervention for young SGM people, which consists of 4 motivational interviewing-based counseling sessions. The YMHP condition will receive YMHP sessions and will be compared to the uTECH+YMHP condition, which includes YMHP sessions as well as uTECH SMS text messages. METHODS: In a study funded by the National Institutes of Health, we will recruit and enroll SGM participants (aged 18-29 years) in the United States (N=330) to participate in a 12-month, 2-arm randomized comparison trial. All participants will receive 4 counseling sessions conducted over Zoom (Zoom Video Communications, Inc) with a master's-level social worker. Participants in the uTECH+YMHP condition will receive curated SMS text messages informed by an ML algorithm that seek to promote HIV and substance use risk reduction strategies as well as undergoing YMHP counseling. We hypothesize that the uTECH+YMHP intervention will be considered acceptable, appropriate, and feasible to most participants. We also hypothesize that participants in the combined condition will experience enhanced and more durable reductions in substance use and sexual risk behaviors compared to participants receiving YMHP alone. Appropriate statistical methods, models, and procedures will be selected to evaluate primary hypotheses and behavioral health outcomes in both intervention conditions using an α<.05 significance level, including comparison tests, tests of fixed effects, and growth curve modeling. RESULTS: This study was funded in August 2019. As of June 2024, all participants have been enrolled. Data analysis has commenced, and expected results will be published in the fall of 2025. CONCLUSIONS: This study aims to develop and test the acceptability, appropriateness, and feasibility of uTECH, a novel approach to reduce HIV risk and substance use among SGM young adults. TRIAL REGISTRATION: ClinicalTrials.gov NCT04710901; https://clinicaltrials.gov/study/NCT04710901. INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): DERR1-10.2196/58448.

Online clustering algorithms.

We introduce a set of clustering algorithms whose performance function is such that the algorithms overcome one of the weaknesses of K-means, its sensitivity to initial conditions which leads it to converge to a local optimum rather than the global optimum. We derive online learning algorithms and illustrate their convergence to optimal solutions which K-means fails to find. We then extend the algorithm by underpinning it with a latent space which enables a topology preserving mapping to be found. We show visualisation results on some standard data sets.

Two forms of immediate reward reinforcement learning for exploratory data analysis.

We review two forms of immediate reward reinforcement learning: in the first of these, the learner is a stochastic node while in the second the individual unit is deterministic but has stochastic synapses. We illustrate the first method on the problem of Independent Component Analysis. Four learning rules have been developed from the second perspective and we investigate the use of these learning rules to perform linear projection techniques such as principal component analysis, exploratory projection pursuit and canonical correlation analysis. The method is very general and simply requires a reward function which is specific to the function we require the unit to perform. We also discuss how the method can be used to learn kernel mappings and conclude by illustrating its use on a topology preserving mapping.

Topological mappings of video and audio data.

We review a new form of self-organizing map which is based on a nonlinear projection of latent points into data space, identical to that performed in the Generative Topographic Mapping (GTM).(1) But whereas the GTM is an extension of a mixture of experts, this model is an extension of a product of experts.(2) We show visualisation and clustering results on a data set composed of video data of lips uttering 5 Korean vowels. Finally we note that we may dispense with the probabilistic underpinnings of the product of experts and derive the same algorithm as a minimisation of mean squared error between the prototypes and the data. This leads us to suggest a new algorithm which incorporates local and global information in the clustering. Both ot the new algorithms achieve better results than the standard Self-Organizing Map.

Controlling the shape and alignment of mesopores by confinement in colloidal crystals: designer pathways to silica monoliths with hierarchical porosity.

Monolithic pieces of hierarchically structured silica, containing both periodic macropores and mesopores with well-controlled architecture, are synthesized by dual templating methods. Colloidal crystal templating with close-packed arrays of poly(methyl methacrylate) spheres yields regular, highly interconnected macropores a few hundred nanometers in diameter, and templating with nonionic surfactants produces mesoporosity (2.5-5.1 nm pore diameters) in the macropore walls. Several distinct mesostructures can be achieved within the silica skeleton, depending on the choice of surfactant, co-surfactant, and processing conditions. In the three-dimensional (3D) confinement of the colloidal crystal template, wormlike channels, cubic (Pm3n), or two-dimensional (2D) hexagonal (P6mm) mesostructures are produced with the surfactant Brij 56 (C16H33(OCH2CH2)nOH (n approximately 10) and dodecane as cosurfactant. In the 2D hexagonal structure, channels are oriented perpendicular to the polymer spheres, thereby connecting adjacent macropores through the silica walls. This orientation contrasts with channel alignment parallel to latex spheres when the polymeric surfactant Pluronic P123 (EO20PO70EO20) is used. On the basis of high-resolution 3D transmission electron microscopy, scanning electron microscopy, small-angle X-ray scattering, and nitrogen sorption measurements, structural and textural properties of the monoliths are described in detail as a function of the synthesis parameters. The control over the mesoarchitecture of these silica-surfactant systems in 3D confinement is explained by considering the relative dimensions of the mesostructures with respect to the interstitial space in the latex template, interfacial interactions, entropic effects, and structural frustration.

Optimization, standardization, and testing of a new NMR method for the determination of zeolite host-organic guest crystal structures.

An optimized and automated protocol for determining the location of guest sorbate molecules in highly siliceous zeolites from (29)Si INADEQUATE and (1)H/(29)Si cross polarization (CP) magic-angle spinning (MAS) NMR experiments is described. With the peaks in the (29)Si MAS NMR spectrum assigned to the unique Si sites in the zeolite framework by a 2D (29)Si INADEQUATE experiment, the location of the sorbate molecule is found by systematically searching for sorbate locations for which the measured rates of (1)H/(29)Si cross polarization of the different Si sites correlate linearly with (1)H/(29)Si second moments calculated from H-Si distances. Due to the (1)H/(29)Si cross polarization being in the "slow CP regime" for many zeolite-sorbate complexes, it is proposed that the CP rate constants are best measured by (1)H/(29)Si cross polarization drain experiments, if possible, to avoid complications that may arise from fast (1)H and (29)Si T(1)rho relaxations. An algorithm for determining the sorbate molecule location is described in detail. A number of ways to effectively summarize and display the large number of solutions which typically result from a prediction of the structure from the CP MAS NMR data are presented, including estimates of the errors involved in the structure determinations. As a working example throughout this paper, the structure of the low loaded p-dichlorobenzene/ZSM-5 complex is determined under different conditions from solid-state (1)H/(29)Si CP MAS NMR data, and the solutions are shown to be in excellent agreement with the known single-crystal X-ray diffraction structure. This structure determination approach is shown to be quite insensitive to the use of relative rate constants rather than absolute values, to the detailed structure of the zeolite framework, and relatively insensitive to temperature and motions.

Solid state NMR method for the determination of 3D zeolite framework/sorbate structures: 1H/29Si CP MAS NMR study of the high-loaded form of p-xylene in ZSM-5 and determination of the unknown structure of the low-loaded form.

A general protocol is described for structure determinations of organic sorbate-zeolite complexes based on the selective, through-space, distance-dependent transfer of magnetization from protons in selectively deuterated organics to framework silicon nuclei. The method was developed using the known structure of the high-loaded ZSM-5/p-xylene complex containing p-xylene-d(6) or p-xylene-d(4). It was then applied to determine the unknown structure of the low-loaded ZSM-5/p-xylene complex using NMR alone. For the high-loaded complex improved data were obtained below 273 K, where slow motions and exchange processes of the p-xylene are eliminated. The general approach was validated by the exact agreement of the experimental (1)H-(29)Si CPMAS spectra obtained at a specific contact time and the complete 24-line spectra simulated using 1/T(CP) vs M(2) correlations from only the six clearly resolved resonances. For the low-loaded complex the (29)Si resonances were assigned at 267 K, and variable contact time CP experiments were carried out between 243 and 173 K using the same specifically deuterated p-xylenes. All possible locations and orientations of the p-xylene guests were sampled, and those solutions that gave acceptable linear 1/T(CP) vs M(2) correlations were selected. The optimum p-xylene location in this temperature range was determined to be in the channel intersection with the long molecular axis parallel to [0,1,0] (ring center fractional coordinates {-0.009, 0.250, 0.541}) with the ring plane oriented at an angle of 30 +/- 3 degrees about the crystallographic b axis. A subsequent single-crystal X-ray study confirmed this predicted structure.

Determining the geometry of strongly hydrogen-bonded silanols in a layered hydrous silicate by solid-state nuclear magnetic resonance.

High-resolution solid-state NMR spectroscopy is exploited to obtain structural constraints involving strongly hydrogen-bonded silanols in octosilicate, a prominent member of the layered hydrous sodium silicates. Proton-silicon cross-polarization dynamics reveals that octosilicate contains two types of Q(3) silicons present in hydrogen-bonded -Si-O-Hcdots, three dots, centeredO-Si- and -Si-O(-)-type sites which can only be distinguished by their different abilities to cross polarize and the different mobilities of neighboring hydrous species. The theoretical analysis of the oscillating components of the polarization transfer buildup curves suggests that the model of heteronuclear pairs is an adequate description of the quantum spin system within hydrogen-bonded -Si-O-Hcdots, three dots, centeredO-Si- fragments. We also show that dipolar modulated, slow speed magic-angle (29)Si NMR spectrum provides unique geometric information on strongly hydrogen-bonded silanols. The dipolar modulated spinning sidebands contain all the information necessary to determine the internuclear Sicdots, three dots, centeredH distances as well as the magnitude and orientation of the principal elements of the (29)Si chemical shielding tensor in the molecular frame. The data provide definite proof of the intralayer character of strongly hydrogen-bonded silanol groups in a bridging, albeit not symmetric, position between neighboring tetrahedra. The approach developed in this work may be useful to obtain structural information on related layered alkali metal silicates, silica gels as well as on other classes of microporous materials.

Symmetry-based 29Si dipolar recoupling magic angle spinning NMR spectroscopy: a new method for investigating three-dimensional structures of zeolite frameworks.

A new 29Si solid-state MAS NMR experiment is described for investigating the framework structures of pure silica zeolites. The symmetry-based homonuclear dipolar recoupling sequence SR26411 has been incorporated into a two-dimensional NMR experiment to probe the Si-O-Si bonding connectivities and long-range Si-Si distances in zeolite frameworks. This dipolar recoupling sequence is shown to have a number of advantages over the J-coupling-based INADEQUATE experiment. For the clathrasil Sigma-2, it is demonstrated that there is excellent agreement between experimental double-quantum build-up curves obtained from a series of two-dimensional double-quantum correlation spectra and simulated curves which consider all Si-Si distances out to 8 A. This result suggests that this experiment could be used to solve zeolite frameworks with unknown structures.

Measurement of NMR cross-polarization (CP) rate constants in the slow CP regime: relevance to structure determinations of zeolite-sorbate and other complexes by cp magic-angle spinning NMR.

When analyzing I --> S variable contact time cross-polarization (CP) curves, the spin dynamics are usually assumed to be describable in the "fast CP regime" in which the growth of the S spin magnetization is governed by the rate of cross polarization while its decay is governed by the rate of I spin T1rho relaxation. However, in the investigation of the structures of zeolite-sorbate and other complexes by polarization transfer this will not necessarily be the case. We discuss the measurement of I --> S CP rate constants under the "slow CP regime" in which the rate of T1rho relaxation is fast compared to the rate of cross polarization, leading to a reversal of the usual assumptions such that the rate or growth is governed by the rate of I spin T1rho relaxation while the decay is governed by the rate of cross polarization (and the S spin T1rho relaxation). It is very important to recognize when a system is in the slow CP regime, as an analysis assuming the normal fast CP will lead to erroneous data. However, even when the slow CP regime is recognized, it is difficult to obtain absolute values for the CP rate constants from fits to standard CP curves, since the CP rate constant is correlated to the scaling factor, the contribution from 29Si T1rho relaxation is ignored, and it is difficult to obtain reliable data at very long contact times. The use of a 29Si{1H} CP "drain" or "depolarization" experiment, which measures absolute values of the CP rate constants, is therefore proposed as being most appropriate for theses situations. To illustrate the importance of these observations, measurements of the 1H-29Si CP rate constants in the p-dichlorobenzene/ZSM-5 sorbate-zeolite complex by 29Si{1H} CP and CP drain magic-angle spinning (MAS) NMR experiments are presented and compared and used to determine the location of the guest sorbate molecules in the cavities of the host zeolite framework.

Zeolite nanoclusters coated onto the mesopore walls of SBA-15.

Ultrahigh-field 27Al MAS and MQMAS NMR and 129Xe NMR were used to characterize the zeolite-coated SBA-15 materials. This study shows evidence that zeolite nanoclusters are coated on the mesopore surface of SBA-15. These techniques are useful for the detection of zeolite nanoclusters and different aluminum environments in zeolite/aluminosilicate composites, which are difficult or even impossible to detect by conventional techniques.

Twinned principal curves.

Principal Curves are extensions of Principal Component Analysis and are smooth curves, which pass through the middle of a data set. We extend the method so that, on pairs of data sets which have underlying non-linear correlations, we have pairs of curves which go through the 'centre' of data sets in such a way that the non-linear correlations between the data sets are captured. The core of the method is to iteratively average the current local projections of the data points which produces an increasingly sparsified set of nodes. The Twinned Principal Curves are generated in three ways: by joining up the nodes in order, by performing Local Canonical Correlation Analysis and by performing Local Exploratory Correlation Analysis (Koetsier et al., 2002). The latter two are shown to improve the forecasting capability of the method but at an increased computational load. We show that it is crucial to terminate the algorithm after a small number of iterations for the first method and investigate several criteria for doing so.

A canonical correlation neural network for multicollinearity and functional data.

We review a recent neural implementation of Canonical Correlation Analysis and show, using ideas suggested by Ridge Regression, how to make the algorithm robust. The network is shown to operate on data sets which exhibit multicollinearity. We develop a second model which not only performs as well on multicollinear data but also on general data sets. This model allows us to vary a single parameter so that the network is capable of performing Partial Least Squares regression (at one extreme) to Canonical Correlation Analysis (at the other)and every intermediate operation between the two. On multicollinear data, the parameter setting is shown to be important but on more general data no particular parameter setting is required. Finally, we develop a second penalty term which acts on such data as a smoother in that the resulting weight vectors are much smoother and more interpretable than the weights without the robustification term. We illustrate our algorithms on both artificial and real data.

Effect of molecular oxygen on the variable-temperature 29Si MAS NMR spectra of zeolite-sorbate complexes.

Structure determinations of siliceous zeolite-sorbate host-guest complexes by solid-state NMR require highly resolved 29Si MAS NMR spectra. As the temperature is lowered, the 29Si MAS NMR spectra of many zeolite-sorbate complexes become broadened such that the resolution of the individual 29Si peaks is lost, limiting the application of solid-state NMR for structure determination. It is shown that the 29Si peak widths are related to the 29Si T2 relaxation times and that the source of the 29Si relaxation and the line broadening is paramagnetic molecular oxygen in the channels of the zeolite. Removal of the oxygen by purging the sample with nitrogen gas leads to a dramatic increase in the resolution of the 29Si MAS NMR spectrum of the p-dibromobenzene/ZSM-5 complex. An analysis of the individual 29Si T1 relaxation times reveals that the oxygen molecules are localized mainly in the zigzag channels of ZSM-5, suggesting that the p-dibromobenzene molecules are located in the channel intersections.

Carbohydrate-binding modules recognize fine substructures of cellulose.

Competition isotherms are used to identify the set of cellulose substructures to which cellulose binding modules (CBMs) from families 2a, 3, 4, 9, and 17 bind. The experiments are based on coupling a unique fluorescent tag to each CBM in a manner that does not alter the natural binding properties of the CBM and therefore allows the surface and solution concentrations of each CBM to be monitored as a function of time and composition. Adsorption and surface exchange of like or competing CBMs are monitored using a range of cellulose preparations varying in both crystallinity and provenance. CBMs from families 2a, 3, 4, 9, and 17 are shown to recognize different physical forms of prepared cellulose. The demonstration of the very fine binding specificity of cellulose-specific CBMs implies that the polysaccharide targets of CBMs extend down to the resolution of cellulose microstructures.

Combined solid state NMR and X-ray diffraction investigation of the local structure of the five-coordinate silicon in fluoride-containing as-synthesized STF zeolite.

The local structure of the [SiO(4/2)F]- unit in fluoride-containing as-synthesized STF zeolite has been experimentally determined by a combination of solid-state NMR and microcrystal X-ray diffraction to be very close to trigonal bipyramidal. Because the fluoride ions are disordered over two sites, the resulting local structure of the [SiO(4/2)F]- unit from a conventional XRD refinement is an average between tetrahedral SiO(4/2) and five-coordinate [[SiO(4/2)F]-, giving an apparent F-Si distance longer than expected. The correct F-Si distance was determined by slow spinning MAS and fast spinning (19)F/(29)Si CP and REDOR solid-state NMR experiments and found to be between 1.72 and 1.79 A. In light of this, the X-ray structure was re-refined, including the disorder at Si3. The resulting local structure of the [SiO(4/2)F]- unit was very close to trigonal bipyramidal with a F-Si distance of 1.744 (6) A, in agreement with the NMR results and the prediction of Density Functional Theory calculations. In addition, further evidence for the existence of a covalent F-Si bond is provided by a (19)F-->(29)Si refocused INEPT experiment. The resonance for the five-coordinate species at -147.5 ppm in the (29)Si spectrum is a doublet due to the (19)F/(29)Si J-coupling of 165 Hz. The peaks in this doublet have remarkably different effective chemical shift anisotropies due to the interplay of the CSA, dipolar coupling, and J-coupling tensors. The distortions from tetrahedral geometry of the neighboring silicon atoms to the five-coordinate Si3 atom are manifested in increased delta(aniso) values. This information, along with F-Si distances measured by (19)F-->(29)Si CP experiments, makes it possible to assign half of the (29)Si resonances to unique tetrahedral sites. As well as determining the local geometry of the [SiO(4/2)F]- unit, the work presented here demonstrates the complementarity of the solid-state NMR and X-ray diffraction techniques and the advantages of using them together.

A Comparative Study of Two Neural Methods of Exploratory Projection Pursuit.

We investigate two recently proposed methods of performing exploratory data analysis using artificial neural networks: the BCM neuron and the negative feedback network. Our findings are that the BCM neuron performs only a severely limited version of Exploratory Projection Pursuit while the negative feedback network is effective in finding all interesting projections of data. The improved properties of an extension to the basic BCM neuron are analysed. Copyright 1997 Elsevier Science Ltd. All Rights Reserved.