We present a framework, which we call Molecule Deep $Q$-Networks (MolDQN), for molecule optimization by combining domain knowledge of chemistry and state-of-the-art reinforcement learning techniques (double $Q$-learning randomized value functions). directly define modifications on molecules, thereby ensuring 100\% chemical validity. Further, operate without pre-training any dataset to avoid possible bias from the choice that set. Inspired problems faced during medicinal lead optimization,...

Deep reinforcement learning was employed to optimize chemical reactions. Our model iteratively records the results of a reaction and chooses new experimental conditions improve outcome. This outperformed state-of-the-art blackbox optimization algorithm by using 71% fewer steps on both simulations real Furthermore, we introduced an efficient exploration strategy drawing from certain probability distributions, which resulted in improvement regret 0.062 0.039 compared with deterministic policy....

Chemical reactions can be greatly accelerated in microdroplets, but the factors that lead to acceleration are still being elucidated. Using rhodamine 6G (R6G) as a model compound, we studied density distribution and fluorescence polarization anisotropy of this dye water-in-oil microdroplets. We found R6G is higher on surface ratio center grows with increasing microdroplet radius or decreasing concentration. The measured at almost same for droplets different sizes becomes larger when...

Desorption electrospray ionization-mass spectrometry imaging (DESI-MSI) was applied to latent fingerprints obtain not only spatial patterns but also chemical maps. Samples with similar lipid compositions as those of the were collected by swiping a glass slide across forehead consenting adults. A machine learning model called gradient boosting tree ensemble (GDBT) samples that allowed us distinguish between different genders, ethnicities, and ages (within 10 years). The results from 194...

Andrea Madotto, Zhaojiang Lin, Zhenpeng Zhou, Seungwhan Moon, Paul Crook, Bing Liu, Zhou Yu, Eunjoon Cho, Pascale Fung, Zhiguang Wang. Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing. 2021.

Zhaojiang Lin, Bing Liu, Seungwhan Moon, Paul Crook, Zhenpeng Zhou, Zhiguang Wang, Zhou Yu, Andrea Madotto, Eunjoon Cho, Rajen Subba. Proceedings of the 2021 Conference North American Chapter Association for Computational Linguistics: Human Language Technologies. 2021.

Zhaojiang Lin, Bing Liu, Andrea Madotto, Seungwhan Moon, Zhenpeng Zhou, Paul Crook, Zhiguang Wang, Zhou Yu, Eunjoon Cho, Rajen Subba, Pascale Fung. Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing. 2021.

Cell-type-specific metabolic profiling in tissue with heterogeneous composition has been of great interest across all mass spectrometry imaging (MSI) technologies. We report here a powerful new chemical capability desorption electrospray ionization (DESI) MSI, which enables cell-type-specific and situ complex samples. accomplish this by combining DESI-MSI immunofluorescence staining using specific cell-type markers. take advantage the variable frequency each distinct cell type lateral septal...

Abstract Artificial control of colors based on plasmonic and dielectric nanostructures is not only interesting for fundamental research, but also important practical applications. Apart from spatial resolution, chromaticity, angular independence, dynamical color‐tuning have emerged as the most desirable functions structural color. Here, it shown theoretically that orientations electric or magnetic dipoles induced in a silicon nanoparticle can be manipulated by using an evanescent wave...

Lipid profile changes in heart muscle have been previously linked to cardiac ischemia and myocardial infarction, but the spatial distribution of lipids metabolites ischemic remains be fully investigated. We performed desorption electrospray ionization mass spectrometry imaging hearts from vivo infarction mouse models. In these mice, was induced by blood supply restriction via a permanent ligation left anterior descending coronary artery. showed that applying machine learning algorithm...

A method called nanotip ambient ionization mass spectrometry (NAIMS) is described, which applies high voltage between a tungsten and metal plate to generate plasma in ionized analytes on the surface of are directed inlet analyzed by spectrometer. The dependence signal intensity investigated as function tip-to-plate distance, tip size, applied at tip, current. These parameters separately optimized achieve sensitivity or spatial resolution. partially observable Markov decision process used...

In this paper, we presented a novel convolutional neural network framework for graph modeling, with the introduction of two new modules specially designed graph-structured data: $k$-th order convolution operator and adaptive filtering module. Importantly, our High-order Adaptive Graph Convolutional Network (HA-GCN) is general-purposed architecture that fits various applications on both node centrics, as well generative models. We conducted extensive experiments demonstrating advantages...

Silver-enhanced fluorescence was coupled with a bio-barcode assay to facilitate dual amplification demonstrate non-enzymatic approach for simple and sensitive detection of DNA. In the design, magnetic nanoparticles seeded silver were modified capture DNA, binding ssDNA fluorescently labeled barcode dsDNA. Upon introduction target sandwich structure formed because hybridization reaction. By separation, silver-enhanced DNAs could be readily measured without need further step liberate from...

The gold standard for cystic fibrosis (CF) diagnosis is the determination of chloride concentration in sweat. Current testing methodology takes up to 3 h complete and has recognized shortcomings on its diagnostic accuracy. We present an alternative method identification CF by combining desorption electrospray ionization mass spectrometry a machine-learning algorithm based gradient boosted decision trees analyze perspiration samples. This process as little 2 min, we determined accuracy be 98...

To develop chemical kinetics models for the combustion of ionic liquid-based monopropellants, identification elementary steps in thermal and catalytic decomposition components such as 2-hydroxyethylhydrazinium nitrate (HEHN) is needed but currently not well understood. The first step protic liquids HEHN typically proton transfer from cation to anion, resulting formation 2-hydroxyethylhydrazine (HEH) HNO3. In part this investigation, high-temperature HEH probed with flash pyrolysis (<1400 K)...

A neutron imaging detector based on neutron-sensitive microchannel plates (MCPS) was constructed and tested at beamlines of thermal cold neutrons. The MCPS are made a glass mixture containing 10B natural Gd, which makes the bulk MCP an efficient converter. Contrary to scintillator screens normally used in imaging, spatial resolution is not traded off with detection efficiency. While best scintillators have efficiency around percent, 50% for neutrons 70% has been demonstrated these earlier....

The conversion of heavy oil fractions into lighter is highly essential for meeting the growing demand fuels in continuous depletion reserves and environmental remediation. We demonstrate an degradation approach where microdroplets are sprayed air at room temperature onto untreated TiO2 nanoparticles-coated filter paper to which we apply electrical potential. Two model polycyclic aromatic hydrocarbons as well a crude sample show on microsecond time scale. Heterogeneous catalysis biphasic...

Abstract End‐stage renal disease (ESRD), characterized by cessation in kidney function, has been linked to severe metabolic disturbances, caused the buildup of toxic solutes blood. To remove these solutes, ESRD patients undergo dialysis. As a proof concept, we tested whether ESRD‐related signatures can be detected perspiration samples using combined methodology. Our rapid methodology involves swabbing glass slide across patient's forehead, detecting metabolites imprint desorption...

End stage renal disease (ESRD), characterized by cessation in kidney function, has been linked to severe metabolic disturbances, caused buildup of toxic solutes blood. To remove these solutes, ESRD patients undergo dialysis. As a proof concept, we tested whether ESRD-related signatures can be detected perspiration samples using combined methodology. Our rapid methodology involves swabbing glass slide across the patient’s forehead, detecting metabolites imprint desorption electrospray...

An amendment to this paper has been published and can be accessed via a link at the top of paper.

Zero-shot cross-domain dialogue state tracking (DST) enables us to handle task-oriented in unseen domains without the expense of collecting in-domain data. In this paper, we propose a slot description enhanced generative approach for zero-shot DST. Specifically, our model first encodes context and slots with pre-trained self-attentive encoder, generates values an auto-regressive manner. addition, incorporate Slot Type Informed Descriptions that capture shared information across facilitate...