National Institute of Science Education and Research,Homi Bhabha National Institute

Abstract This study assesses the impact of physicochemical pretreatment, enzymatic hydrolysis, and co-culture fermentation strategies on bioethanol production from corn husk biomass (CHB). Under optimal alkali pretreatment conditions (1.75% alkali, 4.0 g substrate concentration, 120 °C, 10 h), 35% lignin removal was achieved, with 48% cellulose and 39% hemicellulose recovery. In contrast, acid pretreatment resulted in 30% lignin removal, 45% cellulose recovery, and 34% hemicellulose recovery, showing lower efficiency than alkali pretreatment. During ultrasonication alkali pretreatment enhanced cellulose and hemicellulose exposure up to 51 and 46% and delignification up to 49%. Enzymatic hydrolysis of pretreated corn husk biomass was performed using commercial enzymes [Celluclast 1.5 L (700 EGU or 854 U mL−1) and Viscozyme (13.4 FBG/mL)] and isolated bacterial enzymes, including cellulase from Bacillus licheniformis (9.3 ± 0.3 U mL−1) and xylanase from Enterobacter asburiae PQ396173 (7.0 ± 0.4 U mL−1). The developed enzyme cocktail in ratio 3:2:3:1 (v/v; U mL−1) (Celluclast: Viscozyme: native cellulase: native xylanse) using a cocktail of native and commercial enzymes, yielded total reducing sugar of 740 mg g−1 glucose and 54.6 mg g−1 xylose. Fermentation of hydrolysate prepared with commercial enzymes using monoculture of Saccharomyces cerevisiae and Pichia pastoris yielded 17.6 g L−1 and 12.2 g L−1 bioethanol separately. Co-cultured yeasts produced 26.8 g L−1 ethanol at 96 h of incubation, exceeding monoculture yields. The fermentation with integration of commercial and isolated bacterial enzyme cocktails yielded the highest bioethanol output of 37.3 g L−1 at 96 h incubation, indicating that enzymatic saccharification with a combination of commercial and native enzyme cocktails results in maximum bioethanol production. Graphical abstract

The sensitivity of conventional direct dark matter searches for weakly interacting massive particles (WIMPs) is ultimately limited by coherent elastic neutrino-nucleus scattering (CE nu NS), which produces nuclear recoils indistinguishable from WIMP signals and defines the so-called neutrino floor. While the effects of solar neutrinos, geoneutrinos, diffuse supernova neutrinos, and atmospheric neutrinos have been extensively studied in this context, the contribution from reactor antineutrinos has received comparatively little attention. We present the first systematic evaluation of how reactor antineutrino fluxes, modeled as a function of reactor-detector distance, modify the neutrino floor for low-mass WIMP searches using of the neutrino floor are examined under consistent assumptions. We find that proximity to gigawatt-scale reactors within 10 km can raise the neutrino floor by up to a few orders of magnitude, significantly reducing the sensitivity to sub-10 GeV/c2 dark matter. Beyond 100 km, the reactor contribution becomes negligible. These conclusions hold for both definitions of the neutrino floor and remain stable under reasonable variations in detector quenching, site-dependent geoneutrino flux, and reactor antineutrino flux uncertainties, emphasizing reactor proximity as a critical factor in site selection for future low-threshold dark matter experiments.

The intersection of open-source processor architectures and machine learning is driving the demand for customizable, efficient, and accessible hardware. This survey examines the state of the RISC-V ISA in machine learning applications, analyzing current capabilities, challenges, and future directions based on recent research. The analysis covers academic and commercial implementations, software frameworks, and real-world applications. The RISC-V machine learning ecosystem is evaluated, from instruction set extensions and core implementations to compiler optimizations and deployment strategies. Key contributions include a unified taxonomy of RISC-V ML implementations, a comparative analysis of performance and design trade-offs, an evaluation of software toolchain maturity, and the identification of emerging trends in instruction set extensions and specialized accelerators. Findings reveal progress in energy efficiency, specialized instruction development, and framework integration, while highlighting challenges in standardization, verification complexity, and ecosystem fragmentation. The analysis proposes four research directions to address current limitations: specialized neural processing extensions, adaptive and modular processor architectures, security frameworks, and energy-efficient multi-domain architectures. These directions provide a roadmap for advancing RISC-V as a foundational platform for next-generation machine learning systems.

The spectra of exoplanet atmospheres are affected by aerosols (clouds and hazes) of uncertain origin. Proposed aerosol formation mechanisms include gas condensation or photochemical reactions. We measured the transmission spectrum of the tidally locked gas giant exoplanet WASP-94A b and identified asymmetry in its atmosphere. The morning limb is cooler and cloudy, whereas the evening limb is hotter and exhibits gaseous water absorption features. We interpret this difference as being due to the formation of cloud droplets near the morning limb, which evaporate during circulation to the evening limb. The dominant aerosols are clouds cycling between the day and night sides of the atmosphere, not photochemical hazes. The resulting asymmetry can severely bias chemical abundance measurements, unless limb-resolved spectroscopy is available.

Large Language Models (LLMs) such as ChatGPT are transforming how scientists conduct and validate research, offering promise as tools to improve scientific reproducibility. However, computational reproducibility and error detection remain expensive and labor-intensive. We experimentally test how collaboration between researchers and LLM assistants influences the reproduction of quantitative social science findings across different levels of AI autonomy. We randomly assigned 288 researchers to 103 teams working under three conditions: human-only, AI-assisted (using ChatGPT as a collaborative tool), or AI-led (ChatGPT operating with minimal human oversight). Teams reproduced published results from leading social science journals, detected coding errors, and proposed robustness checks. Human-only and AI-assisted teams achieved comparable reproduction rates (94% vs. 91%) and performed similarly on most outcomes, except human-only teams identified significantly more major coding errors. Both substantially outperformed AI-led teams, which achieved only a 37% reproduction rate, detected fewer errors across all categories, proposed weaker robustness checks, and required more time. This autonomous approach, however, likely represents only a lower bound of AI capabilities. Despite rapid model advances, expert human judgment currently remains indispensable for reliable empirical verification. While AI assistance did not degrade most outcomes, it provided no measurable advantages and was associated with reduced detection of major errors. However, the 37% autonomous reproduction rate indicates that AI could provide value in settings where scale or cost constraints preclude human review of papers, even though general-purpose LLMs offer no immediate advantages for human-supervised verification.