Large language models for scientific discovery in molecular property prediction
Related Articles
A unified acoustic-to-speech-to-language embedding space captures the neural basis of natural language processing in everyday conversations
This study introduces a unified computational framework connecting acoustic, speech and word-level linguistic structures to study the neural basis of everyday conversations in the human brain. We used electrocorticography to record neural signals across 100 h of speech production and comprehension as participants engaged in open-ended real-life conversations. We extracted low-level acoustic, mid-level speech and contextual word embeddings from a multimodal speech-to-text model (Whisper). We developed encoding models that linearly map these embeddings onto brain activity during speech production and comprehension. Remarkably, this model accurately predicts neural activity at each level of the language processing hierarchy across hours of new conversations not used in training the model. The internal processing hierarchy in the model is aligned with the cortical hierarchy for speech and language processing, where sensory and motor regions better align with the model’s speech embeddings, and higher-level language areas better align with the model’s language embeddings. The Whisper model captures the temporal sequence of language-to-speech encoding before word articulation (speech production) and speech-to-language encoding post articulation (speech comprehension). The embeddings learned by this model outperform symbolic models in capturing neural activity supporting natural speech and language. These findings support a paradigm shift towards unified computational models that capture the entire processing hierarchy for speech comprehension and production in real-world conversations.
Language measures correlate with other measures used to study emotion
Researchers are increasingly using language measures to study emotion, yet less is known about whether language relates to other measures often used to study emotion. Building on previous work which focuses on associations between language and self-report, we test associations between language and a broader range of measures (self-report, observer report, facial cues, vocal cues). Furthermore, we examine associations across different dictionaries (LIWC-22, NRC, Lexical Suite, ANEW, VADER) used to estimate valence (i.e., positive versus negative emotion) or discrete emotions (i.e., anger, fear, sadness) in language. Associations were tested in three large, multimodal datasets (Ns = 193–1856; average word count = 316.7–2782.8). Language consistently related to observer report and consistently related to self-report in two of the three datasets. Statistically significant associations between language and facial cues emerged for language measures of valence but not for language measures of discrete emotions. Language did not consistently show significant associations with vocal cues. Results did not tend to significantly vary across dictionaries. The current research suggests that language measures (in particular, language measures of valence) are correlated with a range of other measures used to study emotion. Therefore, researchers may wish to use language to study emotion when other measures are unavailable or impractical for their research question.
A data-driven generative strategy to avoid reward hacking in multi-objective molecular design
Molecular design using data-driven generative models has emerged as a promising technology, impacting various fields such as drug discovery and the development of functional materials. However, this approach is often susceptible to optimization failure due to reward hacking, where prediction models fail to extrapolate, i.e., fail to accurately predict properties for designed molecules that considerably deviate from the training data. While methods for estimating prediction reliability, such as the applicability domain (AD), have been used for mitigating reward hacking, multi-objective optimization makes it challenging. The difficulty arises from the need to determine in advance whether the multiple ADs with some reliability levels overlap in chemical space, and to appropriately adjust the reliability levels for each property prediction. Herein, we propose a reliable design framework to perform multi-objective optimization using generative models while preventing reward hacking. To demonstrate the effectiveness of the proposed framework, we designed candidates for anticancer drugs as a typical example of multi-objective optimization. We successfully designed molecules with high predicted values and reliabilities, including an approved drug. In addition, the reliability levels can be automatically adjusted according to the property prioritization specified by the user without any detailed settings.
Accurately adjusted phenothiazine conformations: reversible conformation transformation at room temperature and self-recoverable stimuli-responsive phosphorescence
Conformational flexibility is essential to the stimuli-responsive property of organic materials, but achieving the reversible molecular transformation is still challenging in functional materials for the high energy barriers and restriction by intermolecular interactions. Herein, through the incorporation of various steric hindrances into phenothiazine derivatives with different positions and quantities to tune the molecular conformations by adjustable repulsive forces, the folded angles gradually changed from 180° to 90° in 17 compounds. When the angle located at 112° with moderated steric effect, dynamic and reversible transformation of conformations under mechanical force has been achieved for the low energy barriers and mutually regulated molecular motions, resulting in both self-recoverable and stimuli-responsive phosphorescence properties for the first time. It opened up a new way to realize the self-recovery property of organic materials, which can facilitate the multi-functional property of smart materials with the opened avenue for other fields with inspiration.
Advancing extrapolative predictions of material properties through learning to learn using extrapolative episodic training
Recent advancements in machine learning have demonstrated its potential to significantly accelerate the discovery of new materials. Central to this progress is the development of rapidly computable property predictors, which allow identifying novel materials with the desired properties from vast material spaces. However, the limited availability of data resources poses a significant challenge in data-driven material research, particularly hindering the exploration of innovative materials beyond the boundaries of existing data. Although machine-learning predictors are inherently interpolative, establishing a general methodology to create an extrapolative predictor remains a fundamental challenge. In this study, we leveraged the attention-based architecture of neural networks and a meta-learning algorithm to enhance extrapolative generalization capabilities. Meta-learners trained repeatedly on arbitrarily generated extrapolative tasks show outstanding generalization for unexplored material spaces. Through the tasks of predicting the physical properties of polymeric materials and hybrid organic–inorganic perovskites, we highlight the potential of such extrapolatively trained models, particularly their ability to rapidly adapt to unseen material domains in transfer-learning scenarios.
Responses