The lead optimization process in drug discovery campaigns is an arduous endeavour where the input of many medicinal chemists is weighed in order to reach a desired molecular property profile. Building the expertise to successfully drive such projects collaboratively is a very time-consuming process that typically spans many years within a chemist's career. In this work we aim to replicate this process by applying artificial intelligence learning-to-rank techniques on feedback that was obtained from 35 chemists at Novartis over the course of several months. We exemplify the usefulness of the learned proxies in routine tasks such as compound prioritization, motif rationalization, and biased de novo drug design. Annotated response data is provided, and developed models and code made available through a permissive open-source license.

Modern science has come to be regarded as an exclusively objective endeavor employing explicit language that attempts to exclude subjective anthropomorphic biases. In doing so it has run the risk of becoming a purely materialistic bias itself, according no proper place to the human spirit or to intuitive insights that have guided the evolution of human cultures, even though this includes the guiding insights of the most important contributors to the sciences. Although this may have been necessary to exclude rampant superstition in the past, a summary overview of the historical evidence indicates a current pressing need to restore a proper balance.

A main challenge in drug discovery is finding molecules with a desirable balance of multiple properties. Here, we focus on the task of molecular optimization, where the goal is to optimize a given starting molecule towards desirable properties. This task can be framed as a machine translation problem in natural language processing, where in our case, a molecule is translated into a molecule with optimized properties based on the SMILES representation. Typically, chemists would use their intuition to suggest chemical transformations for the starting molecule being optimized. A widely used strategy is the concept of matched molecular pairs where two molecules differ by a single transformation. We seek to capture the chemist's intuition from matched molecular pairs using machine translation models. Specifically, the sequence-to-sequence model with attention mechanism, and the Transformer model are employed to generate molecules with desirable properties. As a proof of concept, three ADMET properties are optimized simultaneously: logD, solubility, and clearance, which are important properties of a drug. Since desirable properties often vary from project to project, the user-specified desirable property changes are incorporated into the input as an additional condition together with the starting molecules being optimized. Thus, the models can be guided to generate molecules satisfying the desirable properties. Additionally, we compare the two machine translation models based on the SMILES representation, with a graph-to-graph translation model HierG2G, which has shown the state-of-the-art performance in molecular optimization. Our results show that the Transformer can generate more molecules with desirable properties by making small modifications to the given starting molecules, which can be intuitive to chemists. A further enrichment of diverse molecules can be achieved by using an ensemble of models.

Prior functional imaging studies of moral processing have utilized 'explicit' moral tasks that involve moral deliberation (e.g., reading statements such as 'he shot the victim' and rating the moral appropriateness of the behavior) or 'implicit' moral tasks that involve moral intuition (e.g., reading similar statements and memorizing them for a test but not rating their moral appropriateness). Although the neural mechanisms underlying moral deliberation and moral intuition may differ, these have not been directly compared. Studies using explicit moral tasks have reported increased activity in several regions, most consistently the medial prefrontal cortex and temporo-parietal junction. In the few studies that have utilized implicit moral tasks, medial prefrontal activity has been less consistent, suggesting the medial prefrontal cortex is more critical for moral deliberation than moral intuition. Thus, we hypothesized that medial prefrontal activity would be increased during an explicit, but not an implicit, moral task. Participants (n=28) were scanned using fMRI while viewing 50 unpleasant pictures, half of which depicted moral violations. Half of the participants rated pictures on moral violation severity (explicit task) while the other half indicated whether pictures occurred indoors or outdoors (implicit task). As predicted, participants performing the explicit, but not the implicit, task showed increased ventromedial prefrontal activity while viewing moral pictures. Both groups showed increased temporo-parietal junction activity while viewing moral pictures. These results suggest that the ventromedial prefrontal cortex may contribute more to moral deliberation than moral intuition, whereas the temporo-parietal junction may contribute more to moral intuition than moral deliberation.

Human beings often curb self-interest to develop and enforce social norms, such as fairness, as exemplified in the ultimatum game (UG). Inspired by the dual-system account for the responder's choice during the UG, we investigated whether the neural basis of psychological process induced by fairness is under genetic control using a twin fMRI study (62 monozygotic, 48 dizygotic; mean age: 19.32 ± 1.38 years). We found a moderate genetic contribution to the rejection rate of unfair proposals (24%-35%), independent of stake size or proposer type, during the UG. Using a voxel-level analysis, we found that genetic factors moderately contributed to unfairness-evoked activation in the bilateral anterior insula (AI), regions representing the intuition of fairness norm violations (mean heritability: left 37%, right 40%). No genetic contributions were found in regions related to deliberate, controlled processes in the UG. This study provides the first evidence that evoked brain activity by unfairness in the bilateral AI is influenced by genes and sheds light on the genetic basis of brain processes underlying costly punishment.

We report a methodology using machine learning to capture chemical intuition from a set of (partially) failed attempts to synthesize a metal-organic framework. We define chemical intuition as the collection of unwritten guidelines used by synthetic chemists to find the right synthesis conditions. As (partially) failed experiments usually remain unreported, we have reconstructed a typical track of failed experiments in a successful search for finding the optimal synthesis conditions that yields HKUST-1 with the highest surface area reported to date. We illustrate the importance of quantifying this chemical intuition for the synthesis of novel materials.

Deep neural network (DNN) techniques, as an advanced machine learning framework, have allowed various image diagnoses in plants, which often achieve better prediction performance than human experts in each specific field. Notwithstanding, in plant biology, the application of DNNs is still mostly limited to rapid and effective phenotyping. The recent development of explainable CNN frameworks has allowed visualization of the features in the prediction by a convolutional neural network (CNN), which potentially contributes to the understanding of physiological mechanisms in objective phenotypes. In this study, we propose an integration of explainable CNN and transcriptomic approach to make a physiological interpretation of a fruit internal disorder in persimmon, rapid over-softening. We constructed CNN models to accurately predict the fate to be rapid softening in persimmon cv. Soshu, only with photo images. The explainable CNNs, such as Gradient-weighted Class Activation Mapping (Grad-Class Activation Mapping (CAM)) and guided Grad-CAM, visualized specific featured regions relevant to the prediction of rapid softening, which would correspond to the premonitory symptoms in a fruit. Transcriptomic analyses to compare the featured regions of the predicted rapid-softening and control fruits suggested that rapid softening is triggered by precocious ethylene signal-dependent cell wall modification, despite exhibiting no direct phenotypic changes. Further transcriptomic comparison between the featured and non-featured regions in the predicted rapid-softening fruit suggested that premonitory symptoms reflected hypoxia and the related stress signals finally to induce ethylene signals. These results would provide a good example for the collaboration of image analysis and omics approaches in plant physiology, which uncovered a novel aspect of fruit premonitory reactions in the rapid-softening fate.

Within drug discovery, the goal of AI scientists and cheminformaticians is to help identify molecular starting points that will develop into safe and efficacious drugs while reducing costs, time and failure rates. To achieve this goal, it is crucial to represent molecules in a digital format that makes them machine-readable and facilitates the accurate prediction of properties that drive decision-making. Over the years, molecular representations have evolved from intuitive and human-readable formats to bespoke numerical descriptors and fingerprints, and now to learned representations that capture patterns and salient features across vast chemical spaces. Among these, sequence-based and graph-based representations of small molecules have become highly popular. However, each approach has strengths and weaknesses across dimensions such as generality, computational cost, inversibility for generative applications and interpretability, which can be critical in informing practitioners' decisions. As the drug discovery landscape evolves, opportunities for innovation continue to emerge. These include the creation of molecular representations for high-value, low-data regimes, the distillation of broader biological and chemical knowledge into novel learned representations and the modeling of up-and-coming therapeutic modalities.

Intuition and insight can be deployed on the same continuum. Intuition is the unconscious ability to create links between information; insight is a process by which a sudden comprehension and resolution of a situation arises (i.e. euréka). In the present study, real and virtual environments were used to trigger intuition and insight. The study hypothesised that immersion in real primed environments would facilitate the emergence of intuition and insight in a virtual environment. Forty nine healthy participants were randomly assigned to two groups: "primed" and "non primed." "Primed" participants were immersed in a real environment with olfactory and visual cues; "non primed" participants did not receive any cues. All participants were exposed to a 3D naturalistic virtual environment which represented a district in Paris via a Head Mounted Display (HMD). Locations presented in the virtual scene (i.e. café places) were related to both olfactory and visual primes (i.e. café) and were based on the continuity between real and virtual environments. Once immersed in the virtual environment, all participants were instructed to use their intuition to envision the selected locations during which Skin Conductance Responses (SCRs) and verbal declarations were recorded. When initiation (a) and immersion (b) phases in the virtual environment were considered, "primed" participants had higher SCRs during the immersion phase than the initiation phase in the virtual environment. They showed higher SRCs during the first part of the virtual immersion than "non primed" participants. During the phenomenological interview, "primed" participants reported a higher number of correct intuitive answers than "non primed" participants. Moreover, "primed" participants "with" insight had higher SCRs during real environment immersion than "primed" participants "without" insight. The findings are consistent with the idea that intuitive decisions in various tasks are based on the activation of pre-existing knowledge, which is unconsciously retrieved, but nevertheless can elicit an intuitive impression of coherence and can generate insight.

Clinical intuition and nonanalytic reasoning play a major role in clinical hypothesis generation; however, clinicians' intuition about whether a critically ill child is bacteremic has not been explored. We endeavored to assess pediatric critical care clinicians' ability to predict bacteremia and to evaluate what affected the accuracy of those predictions.

Both intuition and creativity are associated with knowledge creation, yet a clear link between them has not been adequately established. First, the available empirical evidence for an underlying relationship between intuition and creativity is sparse in nature. Further, this evidence is arguable as the concepts are diversely operationalized and the measures adopted are often not validated sufficiently. Combined, these issues make the findings from various studies examining the link between intuition and creativity difficult to replicate. Nevertheless, the role of intuition in creativity should not be neglected as it is often reported to be a core component of the idea generation process, which in conjunction with idea evaluation are crucial phases of creative cognition. We review the prior research findings in respect of idea generation and idea evaluation from the view that intuition can be construed as the gradual accumulation of cues to coherence. Thus, we summarize the literature on what role intuitive processes play in the main stages of the creative problem-solving process and outline a conceptual framework of the interaction between intuition and creativity. Finally, we discuss the main challenges of measuring intuition as well as possible directions for future research.

Computation is increasingly being used to try to accelerate the discovery of new materials. One specific example of this is porous molecular materials, specifically porous organic cages, where the porosity of the materials predominantly comes from the internal cavities of the molecules themselves. The computational discovery of novel structures with useful properties is currently hindered by the difficulty in transitioning from a computational prediction to synthetic realization. Attempts at experimental validation are often time-consuming, expensive, and frequently, the key bottleneck of material discovery. In this work, we developed a computational screening workflow for porous molecules that includes consideration of the synthetic difficulty of material precursors, aimed at easing the transition between computational prediction and experimental realization. We trained a machine learning model by first collecting data on 12,553 molecules categorized either as "easy-to-synthesize" or "difficult-to-synthesize" by expert chemists with years of experience in organic synthesis. We used an approach to address the class imbalance present in our data set, producing a binary classifier able to categorize easy-to-synthesize molecules with few false positives. We then used our model during computational screening for porous organic molecules to bias toward precursors whose easier synthesis requirements would make them promising candidates for experimental realization and material development. We found that even by limiting precursors to those that are easier-to-synthesize, we are still able to identify cages with favorable, and even some rare, properties.

Motivated by the growing fiscal deficits in sub-Saharan Africa, this study examines fiscal deficit's economic, political, and institutional drivers using a panel of twenty-three sub-Saharan African countries. Panel spatial consistent correlation, dynamic fixed effects autoregressive distributed lag, and feasible generalised ordinary least squares were used as the estimation techniques. Our findings reveal that while per capita income, trade openness, population, and religious tension increase the size of fiscal deficit, bureaucracy quality, government stability, Law and order, and military in politics reduce the extent of fiscal deficit. However, corruption control, democratic accountability, and internal conflict have weaker statistical evidence. Furthermore, the study established evidence of long-run co-integration relationships among institutional factors, economic factors, and fiscal deficits in SSA. Per capita income has a significant positive influence in the short run but a negative effect in the long run. Population and religious tension positively impact fiscal deficit in both periods. However, democratic accountability, government stability, and the military in politics significantly negatively impact fiscal deficit in the long run. This study concludes that beyond economic factors, institutional and political factors are significant drivers of fiscal deficit in sub-Saharan Africa. Therefore, strengthening the institutional quality and creating a stable political environment would lessen the accumulation of fiscal deficit.

The study of neural pre-stimulus or "anticipatory" activity opened a new window for understanding how the brain actively constructs the forthcoming reality. Usually, experimental paradigms designed to study anticipatory activity make use of stimuli. The purpose of the present study is to expand the study of neural anticipatory activity upon the temporal occurrence of dichotomic, statistically unpredictable (random) stimuli within an ecological experimental paradigm. To this purpose, we used a simplified driving simulation including two possible, randomly-presented trial types: a car crash end trial and a no car crash end trial. Event Related Potentials (ERP) were extracted -3,000 ms before stimulus onset. We identified a fronto-central negativity starting around 1,000 ms before car crash presentation. By contrast, a whole-scalp distributed positivity characterized the anticipatory activity observed before the end of the trial in the no car crash end condition. The present data are in line with the hypothesis that the brain may also anticipate dichotomic, statistically unpredictable stimuli, relaying onto different pre-stimulus ERP activity. Possible integration with car-smart-systems is also suggested.

Intuition has been defined as the instantaneous, experience-based impression of coherence elicited by cues in the environment. In a context of discovery, intuitive decision-making processes can be conceptualized as occurring within two stages, the first of which comprises an implicit perception of coherence that is not (yet) verbalizable. Through a process of spreading activation, this initially non-conscious perception gradually crosses over a threshold of awareness and thereby becomes explicable. Because of its experiential basis, intuition shares conceptual similarities with implicit memory processes. Based on these, the study addresses two research questions: (1) Is the gradual nature of intuitive processes reflected on a neural level? (2) Do intuition-based decisions differ neurally from priming-based decisions?

In selection decisions, decision makers often struggle to ignore irrelevant information, such as candidates' age, gender and attractiveness, which can lead to suboptimal decisions. One way to correct the effects of these irrelevant attributes is to consider them as suppressor variables, and penalize individuals who unjustifiably benefit from them. Previous research demonstrated that people have difficulties doing so. In five experiments (N = 1325), we examined the mechanism at the core of people's ability to do so. We found that triggering System 2 did not improve participants' ability to correct for this bias. The majority of those who were successful did so even when denied the opportunity to deliberate. We suggest that logic intuition-not deliberation-is the basis for successfully considering irrelevant information as suppressors. Our results are in line with a revised dual-process approach, in which solving reasoning problems can occur directly through System 1 and does not require an override by a System 2's-based process.

Contemporary efforts for empirically-unbiased modeling of protein-ligand interactions entail a painful tradeoff - as reliable information on both noncovalent binding factors and the dynamic behavior of a protein-ligand complex is often beyond practical limits. We demonstrate that information drawn exclusively from static molecular structures can be used for reproducing and predicting experimentally-measured binding affinities for protein-ligand complexes. In particular, inhibition constants (Ki) were calculated for seven different competitive inhibitors of Torpedo californica acetylcholinesterase using a multiple-linear-regression-based model. The latter, incorporating five independent variables - drawn from QM cluster, DLPNO-CCSD(T) calculations and LED analyses on the seven complexes, each containing active amino-acid residues found within interacting distance (3.5 Å) from the corresponding ligand - is shown to recover 99.9% of the sum of squares for measured Ki values, while having no statistically-significant residual errors. Despite being fitted to a small number of data points, leave-one-out cross-validation statistics suggest that it possesses surprising predictive value (Q2LOO=0.78, or 0.91 upon removal of a single outlier). This thus challenges ligand-invariant definitions of active sites, such as implied in the lock-key binding theory, as well as in alternatives highlighting shape-complementarity without taking electronic effects into account. Broader implications of the current work are discussed in dedicated appendices.

It is well established that religiosity correlates inversely with intelligence. A prominent hypothesis states that this correlation reflects behavioral biases toward intuitive problem solving, which causes errors when intuition conflicts with reasoning. We tested predictions of this hypothesis by analyzing data from two large-scale Internet-cohort studies (combined N = 63,235). We report that atheists surpass religious individuals in terms of reasoning but not working-memory performance. The religiosity effect is robust across sociodemographic factors including age, education and country of origin. It varies significantly across religions and this co-occurs with substantial cross-group differences in religious dogmatism. Critically, the religiosity effect is strongest for tasks that explicitly manipulate conflict; more specifically, atheists outperform the most dogmatic religious group by a substantial margin (0.6 standard deviations) during a color-word conflict task but not during a challenging matrix-reasoning task. These results support the hypothesis that behavioral biases rather than impaired general intelligence underlie the religiosity effect.

No abstract available

People can intuitively estimate the semantic coherence of word triads, even when they are unable to state the triads' common denominator. The present research examines the role of working memory in such intuitive coherence judgments. Dual-process models of information processing suggest that intuition does not depend on working memory. Consistent with this, the authors predicted that taxing working memory capacity will not lower the accuracy of intuitive coherence judgments. Nevertheless, taxing working memory may impede metacognitive processing, which may lead people to become more conservative in judging triads as coherent. Two studies (combined N = 307) tested these predictions by asking participants to memorize letter-number combinations of varying lengths while providing intuitive coherence judgments. As expected, working memory load had no effect on the accuracy of intuitive coherence judgments (Studies 1 & 2). Effects on judgment bias were less consistent. In Study 1, participants became slightly more conservative in judging triads as coherent under moderate (compared to low) working memory load. In Study 2, which was preregistered, working memory load led to more conservative intuitive coherence judgments, but only when participants prioritized a highly demanding load task. Unexpectedly, when focusing on a moderate (compared to a low) working memory load, participants were more liberal in judging triads as coherent. Together, these findings indicate that taxing working memory may interfere with people's inclination to trust their intuition, even when it leaves the accuracy of people's intuition intact.