arXiv subjects

Joy Datta

Publications and source records attributed to Joy Datta.

4 recordsLinked to original sources

arXiv2026

Minimizing the Effect of Sleep Deprivation in the Forward-Forward Algorithm

This paper addresses the challenge posed by sleep deprivation in the Forward-Forward algorithm, where separating the two passes in this algorithm and imbalancing the data processing in the passes is considered an imitation of the cognitive processes observed in humans suffering from sleep deprivation. Previous research has demonstrated that sleep deprivation in the Forward-Forward algorithm has a catastrophic effect on learning efficacy. To mitigate this issue, we explore several approaches; these include alternative activation, optimized loss function, and threshold tuning. To simulate periodic rest, we reduce the number of positive passes in alternating epochs, creating short break phases. We additionally investigate the potential of caffeine-induced stimulation to enhance performance during sleep-deprived conditions. Experimental evaluations conducted on the MNIST and Fashion-MNIST datasets demonstrate that these modifications improve accuracy under the context of sleep deprivation. For example, a 2%-62% accuracy gain is observed in a severe sleep deprivation setting (16 positive or awake periods and 1 negative or sleep period). The approaches also enhance the resilience of the algorithm and its alignment with the adaptive mechanisms of human cognition.

cs.LG↗

arXiv2024

Generative AI for Discovering Porous Oxide Materials for Next-Generation Energy Storage

The key challenge in advancing multivalent-ion batteries lies in finding suitable intercalation hosts. Open-tunnel oxides, featuring one-dimensional channels or nanopores, show promise for enabling effective ion transport. However, the vast range of compositional possibilities renders traditional experimental and quantum-based methods impractical for large-scale studies. This work presents a generative AI framework that uses the Crystal Diffusion Variational Autoencoder (CDVAE) and a fine-tuned Large Language Model (LLM) to expedite the discovery of stable open-tunneled oxide materials for multivalent-ion batteries. By combining machine learning with data mining techniques, five promising transition metal oxide (TMO) structures are generated. These structures, known for forming open-tunnel oxide frameworks, are structurally validated through Density Functional Theory (DFT). The results show that the generated structures have lower formation energies compared to similar compositions in the Materials Project (MP) database, indicating improved thermodynamic stability. Additionally, the graph-based M3GNet model is employed to relax further generated structures, providing a more computationally efficient alternative to DFT. Machine learning-based predictions of formation energy, band gap, and energy above the hull refine the selection process, leading to the identification of materials with significant potential for real-world battery applications. This research demonstrates the power of generative AI in rapidly exploring the vast chemical space of TMOs, offering a new approach to discovering stable open-tunnel oxides for multivalent-ion batteries. The results highlight the potential of this approach to contribute to more sustainable energy storage technologies, addressing the growing concerns surrounding the scarcity of lithium.

cond-mat.mtrl-sci↗

arXiv2023

Transferable and Robust Machine Learning Model for Predicting Stability of Si Anodes for Multivalent Cation Batteries

Data-driven methodology has become a key tool in computationally predicting material properties. Currently, these techniques are priced high due to computational requirements for generating sufficient training data for high-precision machine learning models. In this study, we present a Support Vector Regression (SVR)-based machine learning model to predict the stability of silicon (Si) - alkaline metal alloys, with a strong emphasis on the transferability of the model to new silicon alloys with different electronic configurations and structures. We elaborate on the role of the structural descriptor in imparting transferability to the model that is trained on limited data (~750 Si alloys) derived from the Material Project database. Three popular descriptors, namely X-Ray Diffraction (XRD), Sine Coulomb Matrix (SCM), and Orbital Field Matrix (OFM), are evaluated for representing Si alloys. The material structures are represented by descriptors in the SVR model, coupled with hyperparameter tuning techniques like Grid Search CV and Bayesian Optimization (BO), to find the best performing model for predicting total energy, formation energy and packing fraction of the Si alloy systems. The models are trained on Si alloys with lithium (Li), sodium (Na), potassium (K), magnesium (Mg), calcium (Ca), and aluminum (Al) metals, where Si-Na and Si-Al systems are used as test structures. Our results show that XRD, an experimentally derived characterization of structures, performs most reliably as a descriptor for total energy prediction of new Si alloys. The study demonstrates that by qualitatively selection of training data, using hyperparameter tuning methods, and employing appropriate structural descriptors, the data requirements for robust and accurate ML models can be reduced.

cond-mat.mtrl-sci↗

arXiv2023

Open-tunneled oxides as intercalation host for multivalent ion (Ca and Al) batteries: A DFT study

Lithium-ion batteries (LIBs) are ubiquitous in everyday applications. However, Lithium (Li) is a limited resource on the planet and is therefore not sustainable. As an alternative to lithium, earth-abundant and cheaper multivalent metals such as aluminum (Al) and calcium (Ca) have been actively researched in battery systems. However, finding suitable intercalation hosts for multivalent-ion batteries is urgently needed. Open-tunneled oxides are a particular category of microparticles distinguished by the presence of integrated one-dimensional channels or nanopores. This work focuses on two promising open-tunnel oxides, viz: Niobium Tungsten Oxide (NTO) and Molybdenum Vanadium Oxide (MoVO). We find that the MoVO structure can adsorb greater numbers of multivalent ions than NTO due to its larger surface area and different shapes. The MoVO structure can adsorb Ca, Li, and Al ions with adsorption potential at around 4 to 5 eV. However, the adsorption potential for hexagonal channels of Al ion drops to 1.73 eV because of less channel area. NTO structure has an insertion/adsorption potential of 4.4 eV, 3.4 eV, and 0.9 eV for one Li, Ca, and Al, respectively. In general, Ca ion is more adsorbable than Al ion in both MoVO and NTO structures. Bader charge analysis and charge density plot reveals the role of charge transfer and ion size on the insertion of multivalent ions such as Ca and Al into MoVO and NTO systems. Our results provide general guidelines to explore other multivalent ions for battery applications.

cond-mat.mtrl-sci↗