arXiv subjects

Adarsh Kumar

Publications and source records attributed to Adarsh Kumar.

At least 19 recordsLinked to original sources

arXiv2026

Decomposition of disk and jet optical emission: Case of PKS~1510-089 and PKS~0736+017

Measuring BLR size and black hole mass of jetted quasars with strong radio emission is challenging due to the contribution from non-thermal synchrotron radiation from the jet in the optical continuum. Here, we investigate the optical continuum emission in two flat-spectrum radio quasars (FSRQs), PKS~0736+017 and PKS~1510-089, by separating the thermal emission from the accretion disk and the non-thermal synchrotron emission from the jet using their distinct polarization signatures. Our analysis combines multi-epoch optical spectro-photometric and polarimetric observations from the archival optical continuum, H$β$ emission line light curves from earlier reverberation mapping campaigns, and 37 GHz radio data. For PKS~0736+017, we find that the disk contributes approximately 70 % ($f_d = 0.70^{+0.01}_{-0.01}$) of the total optical emission. After removing the jet component, we obtain a revised H$β$ lag of $84.8^{+23.2}_{-18.4}$ days. In contrast, the disk contribution in PKS~1510-089 is lower with 37.5 % ($f_d = 0.37^{+0.01}_{-0.01}$), with a corresponding H$β$ lag of $54.4^{+4.4}_{-5.4}$ days. Our findings highlight the need for careful decomposition of optical light curves in jet-dominated AGN.

astro-ph.GA↗

arXiv2026

SHIFT-JEV: Certified Abstention and Gain-Gated Continual Adaptation for LLM-Based Decision Models under Distribution Shift

Large language models (LLMs) can act as decision models that return a probability distribution over a fixed set of options, but deployed decision models face domain, label-prior, emerging-class and candidate-set shift. We study how such models fail under shift and propose SHIFT-Jev, a monitoring, abstention and continual-adaptation procedure for them. It scores the risk of each decision from uncertainty, format and option-consistency signals, certifies abstention thresholds with Learn-then-Test, monitors the decision stream with drift and label-shift detectors, and accepts a model update only when an expected-gain gate finds a significant held-out improvement without loss on earlier regimes. On a benchmark of ten shifted test environments with Qwen3-0.6B and Qwen3-4B (three seeds each) and Phi-4-mini, fine-tuning with KL anchors and temperature scaling lowers calibration error under shift from 0.289 to 0.087. At a target risk of 10%, a learned uncertainty score with a certified threshold exceeds the target in 3 to 5 of 27 environment-seed tests, against 5 to 15 for a max-probability score with the same certification and 9 to 14 for conformal prediction sets; drift-based routing changes little. A chi-squared test on predicted labels detects 95 to 99% of class-prior shifts that confidence- and embedding-based detectors miss. In continual adaptation, SHIFT-Jev reaches 70.5% accuracy on Qwen3-0.6B, above replay, LwF, EWC and fine-tuning baselines in every seed, whereas on Qwen3-4B its conservative gate leaves it 0.4 to 1.4 points behind them.

cs.LG↗

arXiv2026

The dawn of the first galaxies

Understanding how the first galaxies assembled their stellar mass at cosmic dawn is a central frontier of modern astrophysics. JWST has now revealed an unexpected population of ultraviolet-bright galaxies at z>10, whose intense emission points towards a rapid stellar mass growth that challenges current models of early galaxy assembly. Yet, the cold gas fueling this growth remains hidden: current ALMA sensitivity restricts routine detection to only the brightest ionized-gas tracers such as [OIII] 88$μ$m, while [CII] 158$μ$m, a key coolant and gas-mass proxy of the neutral interstellar medium (ISM), is undetected beyond $z\sim8$. This leaves fundamental questions unanswered: how quickly primeval galaxies gathered cold gas and turned it into stars; what physical conditions (density, temperature, radiation field) characterized their ISM; how fast metals and dust built up; and how feedback regulated early gas assembly. Only a transformational sub-millimeter facility like ALMA2040 can address these questions, enabling systematic surveys of [CII] in the first galaxies and providing a representative census of gas fractions and depletion times. A simulated galaxy at $z=11.3$ shows that ALMA2040 not only will detect, but also spatially resolve its merging cold-gas structure in [CII] on $\sim$200 pc scales in 5 hours (below 100 pc in 10 hours). ALMA2040 will also spectrally resolve multiple lines simultaneously, akin to integral-field spectroscopy, providing the key diagnostics needed to break density-metallicity-radiation-field degeneracies and map the multi-phase ISM of z>10 galaxies. The combination of dramatically increased line sensitivity and wide instantaneous bandwidth will push the search for the first galaxies out to the earliest cosmic times, with [OIII] detectable up to $z\sim20$. ALMA2040 will thus be transformative for understanding how the first galaxies assembled at cosmic dawn.

astro-ph.IM↗

arXiv2026

Dust at cosmic dawn: unveiling the onset of dust and obscured star formation at $z\gtrsim8$

Although less than 1 percent of the baryonic mass budget of a galaxy, cosmic dust profoundly alters the multi-wavelength appearance of galaxies across cosmic time. The (sub)millimeter wavelength regime is uniquely capable of weighing these dust reservoirs, and revealing the total extent of galaxy evolution that remains hidden behind them. With current facilities, however, doing so at $z > 8$ is largely out of reach. ALMA2040, a next-generation (sub)millimeter interferometer, is set to change this. Through an order-of-magnitude increase in sensitivity compared to ALMA, ALMA2040 is a transformational facility that will be able to routinely detect dust emission from galaxies at $z > 8$ and, through simultaneous multi-band characterization, accurately measure their dust masses, temperatures, and infrared luminosities. Combined with ancillary metallicity measurements -- including from the far-infrared lines detected by ALMA2040 itself -- this will establish the relative roles of supernovae and grain growth in building up dust in the earliest galaxies. Simultaneously, ALMA2040 will measure the fraction of star formation that is obscured by dust at cosmic dawn, and thereby the extent to which rest-UV and optical surveys miss early galaxy build-up. Without a 10$\times$ upgrade to ALMA -- comparable to the leap from the VLT to the ELT -- we will not be able to capitalize on the momentous advances enabled by JWST, and will remain blind to the cold and obscured component of galaxy formation in the $z > 8$ Universe.

astro-ph.IM↗

arXiv2026

Probing cold gas and dust in quiescent galaxies across cosmic time

The discovery of massive quiescent galaxies at high redshift by JWST has revealed that star formation can cease rapidly in the early Universe, yet the physical mechanisms responsible for quenching remain unknown. A key missing element is the systematic measurement in statistically representative quiescent samples of the cold gas and dust content, which are crucial for distinguishing quenching mechanisms and evolutionary pathways. Such observations are currently not feasible because state-of-the-art sub-millimeter facilities (e.g., ALMA) cannot routinely detect the faint molecular gas and dust reservoirs of quiescent galaxies, nor pinpoint outflowing molecular gas. A next-generation facility (ALMA2040) capable of providing transformative access to the cold interstellar medium of quiescent galaxies across redshift is therefore essential. In particular, it will allow the community to probe the molecular gas content of quiescent galaxies at cosmic noon and beyond by targeting the CO, [CI], and [CII] emission lines, as well as unveiling the physical conditions of the cold interstellar medium by constraining the CO spectral line energy distribution. It will also reveal the dust content of the passive population, a crucial ingredient to constrain the different quenching mechanisms. Finally, it will enable probing the molecular gas kinematics and identify molecular outflows driven by supermassive black holes. When combined with the rich legacy of JWST, ELT, and SKAO surveys, ALMA2040 will deliver a comprehensive, multi-phase view of feedback and quenching, allowing us to establish when, how, and why galaxies shut down their star formation across cosmic time.

astro-ph.IM↗

arXiv2026

ALMA2040: Mapping the Spatially Resolved Dust and Star Formation Physics Driving High-z Galaxy Assembly

Interstellar dust reprocesses approximately half of the starlight emitted across cosmic history. However, dust temperature, which is essential to convert continuum flux into dust mass, molecular gas mass, and obscured star-formation rates, remains largely unconstrained on sub-galactic scales at high redshift. While single-band ALMA imaging has characterized continuum sizes for hundreds of z>1 galaxies, single-frequency measurements cannot separate the dust mass column from the local dust heating. To date, spatially resolved (<0.5-arcsec) multi-band dust temperatures exist in the literature for only ~10 extreme systems (quasars, hyperluminous starbursts, or strongly lensed galaxies) at z>1, with ongoing programs within a few years adding at most a few dozen submillimeter or UV-bright galaxies. This still leaves typical main-sequence disks largely unexplored across cosmic history. With an anticipated ~30 times improvement in continuum sensitivity (compared to the current ALMA), simultaneous multi-band receivers, and (sub-)kpc angular resolution, ALMA2040 will transform resolved dust thermometry into a systematic survey science. Here, we present the scientific rationale and technical requirements, highlighting a future statistically meaningful survey with > 100 representative main-sequence galaxies across cosmic epochs. Realistic visibility simulations demonstrate that ALMA2040 measures radial temperature gradients with ~0.5-4 K precision (L_IR~10^{12} L_sun in 2 h on-source at z~2). These measurements will break the degeneracy between intrinsic compactness and temperature gradients, deliver the first extinction-free, temperature-calibrated resolved star-formation relations, and uniquely trace radiative feedback and disk maturation from cosmic noon to cosmic dawn.

astro-ph.IM↗

arXiv2026

Probing ordered magnetic fields in galaxies across cosmic time

Magnetic fields are dynamically important across many astronomical environments, yet their origin, strength, and evolution over cosmic time remain poorly understood. Ordered magnetic fields are ubiquitous in the nearby Universe, as revealed by a range of observational techniques, but evidence for their existence at cosmological distances is scarce. Recent ALMA detections of polarised dust emission indicate that galactic-scale ordered magnetic fields, spanning up to kiloparsec scales, were already present in massive star-forming galaxies at redshifts up to $z\sim5.6$. This white paper argues that the ALMA2040 concept -- with its nearly tenfold improvement in sensitivity and angular resolution, and full-polarisation capability extended to higher-frequency bands -- will open up the study of cosmic magnetism in the early Universe. We identify four science drivers: constraining the dynamos that generate galactic-scale ordered fields, measuring the strength of early magnetic fields, connecting magnetic fields to galaxy evolution across cosmic time, and constraining dust grain properties through the polarised spectral energy distribution. We summarise the corresponding requirements in angular resolution, spectral coverage, and sensitivity, and discuss synergies with next-generation telescopes.

astro-ph.IM↗

arXiv2026

Has JWST Seen Everything? Hunting for "NIRCam-dark" Dusty Star-Forming Galaxies at $z \simeq 6 - 9$ with ALMA2040

JWST/NIRCam has recovered the stellar counterparts of many ALMA-detected, formerly HST-dark galaxies, yet ALMA has started identifying dusty star-forming galaxies (DSFGs) at $z>6$ that remain extremely faint or undetected even with NIRCam. These "NIRCam-dark" DSFGs may trace a missing rapid-growth phase connecting UV-luminous galaxies at $z>10$ to massive quiescent galaxies at $z\gtrsim4$. So far, the small, heterogeneously selected NIRCam-dark sample is insufficient to determine their abundance or role in the evolution of massive galaxies. We show that a blind 1.3-mm continuum survey with ALMA2040 can provide the first high-significance census of this elusive NIRCam-dark DSFG population across $z\simeq 6 - 9$. We conduct sensitivity and survey-speed calculations requiring a $10σ$ detection of a NIRCam-dark DSFG with a dust-obscured star-formation rate of $60\,M_\odot\,{\rm yr}^{-1}$ at $z = 8$. At a like-for-like 7.5-GHz normalization, ALMA2040 will reach the required 1.3-mm RMS of $21\,μ$Jy beam$^{-1}$ about 90 times faster than current ALMA. With 100 hr on-source, ALMA2040 can cover an area of $\sim0.5$ deg$^2$ and detect $\langle N \rangle\sim16$ NIRCam-dark DSFGs over $7.5 10$ and the rapid emergence of massive quiescent systems at $z\gtrsim 4$.

astro-ph.IM↗

arXiv2026

What fuels the growth of stars and black holes in galaxies through cosmic time? The physics and chemistry of the interstellar medium in distant galaxies

The interplay between the baryonic matter in galaxies (gas, dust, stars) has a fundamental role in our understanding of galaxy formation and evolution through cosmic time. The (sub)mm regime provides a unique window to measure the physical conditions and chemistry in the cold and dusty gas reservoirs that fuel the formation of new stars and the growth of massive black holes in distant galaxies. Large sub(mm) interferometers have made a revolutionary impact mapping the cold gas and dust in galaxies beyond the local universe. However, significant progress is hampered by limited spectral-line sensitivity. After more than a decade of surveys, molecular gas (traced by 12CO) has been measured only in a few tens of typical, massive main-sequence star-forming galaxies at z>1, while access to fainter gas-tracing species has been limited to only the very brightest galaxies. Thereby it restricts access to the detailed conditions in the cold gas clouds in typical galaxies in the distant universe. A revolutionary, flagship sub-mm interferometer with a massive increase in instantaneous spectral line sensitivity and broad spectral grasp (ALMA2040) fills this gap by enabling rich, multi-transition, multi-species and multi-isotopologue studies of the atomic and molecular interstellar medium for thousands of typical galaxies across galaxy populations and redshift. This will unveil how the ISM evolves across the galaxy population for statistical samples in different redshift bins, detailed studies of the physical conditions and chemistry in the multi-phase interstellar medium of typical galaxies, and how the cold ISM regulates and responds to star- and black-hole formation, providing a tremendous leap forward in the study of galaxy assembly over cosmic time.

astro-ph.IM↗

arXiv2026

Unveiling the Baryon Cycle Through Cosmic Time

Galaxy evolution is governed by the baryon cycle: gas accreted from the cosmic web and circumgalactic medium condenses into molecular gas, forms stars, and is redistributed by stellar and AGN feedback. Tracing the molecular phase is therefore essential to understanding how galaxies acquire, consume, and eject their baryons. Yet current molecular-gas surveys remain limited by sensitivity, area, cosmic variance, and spectral bandwidth. ALMA2040 would combine a $\sim10\times$ sensitivity improvement with $\gtrsim512$ GHz simultaneous frequency coverage, delivering $\sim100\times$ faster mapping at fixed depth and a further $\gtrsim25\times$ gain in instantaneous bandwidth. Together, these advances would increase spectroscopic survey speed by orders of magnitude, enabling tiered ``wedding-cake'' strategies: an ultra-deep $\sim4$-5 arcmin$^2$ field probing faint and high-redshift reservoirs, a $\sim$0.25 deg$^2$ survey reaching molecular-line sensitivities comparable to current ultra-deep fields over a much larger cosmological volume, and thousands of independent pointings minimizing cosmic variance. Broad spectral coverage would simultaneously detect multiple CO transitions, [CI], H$_2$O, feedback tracers, far-IR fine-structure lines, and dust continuum, constraining molecular gas, ISM conditions, obscured star formation, and gas flows within the same observations. Every targeted observation would also become a molecular deep field, extending these measurements to serendipitous galaxies. Together with future HI, optical, and near-IR surveys, ALMA2040 would enable a population-resolved view of the baryon cycle across cosmic time and stringent tests of galaxy formation models.

astro-ph.IM↗

arXiv2025

The PAU Survey: Uncovering the connection between intrinsic and observed galaxy properties using symbolic regression

Estimating stellar masses for billions of galaxies in upcoming surveys requires methods that are both accurate and computationally efficient. We present a new approach using symbolic regression trained on a simulation to derive simple, explicit mathematical expressions that estimate galaxy stellar masses from basic observables: photometry and redshift. Using a mock catalogue from the GALFORM semi-analytical model that reproduces the Physics of the Accelerating Universe Survey (PAUS), we show that a linear combination of just four observables -- minimally processed $u$- and $i$- band magnitudes, observed $(g-r)$ colour, and redshift -- can recover stellar masses with accuracy comparable to traditional spectral energy distribution (SED) fitting, but with negligible computational cost. Our expressions can be evaluated instantaneously for millions of galaxies, making them ideal for next-generation surveys like LSST and Euclid. When observational errors are included, symbolic regression achieves a similar accuracy to deep neural networks while maintaining transparency. Validation against CIGALE SED fitting on PAUS data shows agreement within 0.13 dex for galaxies with $M_{*} > 10^8 M_{\odot}$. We demonstrate that the stellar mass function can be recovered at $z < 0.5$, though with distortions at the extremes: the high-mass end is overestimated by a factor of $\sim 3$ at $10^{11.5} h^{-1} M_{\odot}$ due to scatter. Our approach offers a fast, transparent alternative to traditional methods without sacrificing accuracy for the bulk of the galaxy population.

astro-ph.GA↗

arXiv2025

Multimodal Fusion of Glucose Monitoring and Food Imagery for Caloric Content Prediction

Effective dietary monitoring is critical for managing Type 2 diabetes, yet accurately estimating caloric intake remains a major challenge. While continuous glucose monitors (CGMs) offer valuable physiological data, they often fall short in capturing the full nutritional profile of meals due to inter-individual and meal-specific variability. In this work, we introduce a multimodal deep learning framework that jointly leverages CGM time-series data, Demographic/Microbiome, and pre-meal food images to enhance caloric estimation. Our model utilizes attention based encoding and a convolutional feature extraction for meal imagery, multi-layer perceptrons for CGM and Microbiome data followed by a late fusion strategy for joint reasoning. We evaluate our approach on a curated dataset of over 40 participants, incorporating synchronized CGM, Demographic and Microbiome data and meal photographs with standardized caloric labels. Our model achieves a Root Mean Squared Relative Error (RMSRE) of 0.2544, outperforming the baselines models by over 50%. These findings demonstrate the potential of multimodal sensing to improve automated dietary assessment tools for chronic disease management.

cs.CV↗

arXiv2025

Improving the Reliability of LLMs: Combining CoT, RAG, Self-Consistency, and Self-Verification

Hallucination, where large language models (LLMs) generate confident but incorrect or irrelevant information, remains a key limitation in their application to complex, open-ended tasks. Chain-of-thought (CoT) prompting has emerged as a promising method for improving multistep reasoning by guiding models through intermediate steps. However, CoT alone does not fully address the hallucination problem. In this work, we investigate how combining CoT with retrieval-augmented generation (RAG), as well as applying self-consistency and self-verification strategies, can reduce hallucinations and improve factual accuracy. By incorporating external knowledge sources during reasoning and enabling models to verify or revise their own outputs, we aim to generate more accurate and coherent responses. We present a comparative evaluation of baseline LLMs against CoT, CoT+RAG, self-consistency, and self-verification techniques. Our results highlight the effectiveness of each method and identify the most robust approach for minimizing hallucinations while preserving fluency and reasoning depth.

cs.AI↗

arXiv2024

Revolutionizing Modern Networks: Advances in AI, Machine Learning, and Blockchain for Quantum Satellites and UAV-based Communication

Quantum communication is the most secure technique of transmitting data available today. Fiber communication lines and satellite-to-ground links have served as the basis for the most successful quantum networks that have been developed so far. Using a UAV, satellite or both for free-space quantum communication reduces the need for permanent ground connections and takes advantage of the lower loss limit in space, which makes it more efficient. This work surveys the recent development in Quantum Satellites and Quantum UAVs-based networks. Here, the importance of the latest technologies, including quantum artificial intelligence, blockchain quantum machine learning, quantum satellites and quantum UAVs, are explored from network perspectives. Further, this work discussed the role of satellite-based images and artificial intelligence.

cs.NI↗

arXiv2023

ShredGP: Guitarist Style-Conditioned Tablature Generation

GuitarPro format tablatures are a type of digital music notation that encapsulates information about guitar playing techniques and fingerings. We introduce ShredGP, a GuitarPro tablature generative Transformer-based model conditioned to imitate the style of four distinct iconic electric guitarists. In order to assess the idiosyncrasies of each guitar player, we adopt a computational musicology methodology by analysing features computed from the tokens yielded by the DadaGP encoding scheme. Statistical analyses of the features evidence significant differences between the four guitarists. We trained two variants of the ShredGP model, one using a multi-instrument corpus, the other using solo guitar data. We present a BERT-based model for guitar player classification and use it to evaluate the generated examples. Overall, results from the classifier show that ShredGP is able to generate content congruent with the style of the targeted guitar player. Finally, we reflect on prospective applications for ShredGP for human-AI music interaction.

cs.SD↗

arXiv2023

From Words to Music: A Study of Subword Tokenization Techniques in Symbolic Music Generation

Subword tokenization has been widely successful in text-based natural language processing (NLP) tasks with Transformer-based models. As Transformer models become increasingly popular in symbolic music-related studies, it is imperative to investigate the efficacy of subword tokenization in the symbolic music domain. In this paper, we explore subword tokenization techniques, such as byte-pair encoding (BPE), in symbolic music generation and its impact on the overall structure of generated songs. Our experiments are based on three types of MIDI datasets: single track-melody only, multi-track with a single instrument, and multi-track and multi-instrument. We apply subword tokenization on post-musical tokenization schemes and find that it enables the generation of longer songs at the same time and improves the overall structure of the generated music in terms of objective metrics like structure indicator (SI), Pitch Class Entropy, etc. We also compare two subword tokenization methods, BPE and Unigram, and observe that both methods lead to consistent improvements. Our study suggests that subword tokenization is a promising technique for symbolic music generation and may have broader implications for music composition, particularly in cases involving complex data such as multi-track songs.

cs.SD↗

arXiv2023

GTR-CTRL: Instrument and Genre Conditioning for Guitar-Focused Music Generation with Transformers

Recently, symbolic music generation with deep learning techniques has witnessed steady improvements. Most works on this topic focus on MIDI representations, but less attention has been paid to symbolic music generation using guitar tablatures (tabs) which can be used to encode multiple instruments. Tabs include information on expressive techniques and fingerings for fretted string instruments in addition to rhythm and pitch. In this work, we use the DadaGP dataset for guitar tab music generation, a corpus of over 26k songs in GuitarPro and token formats. We introduce methods to condition a Transformer-XL deep learning model to generate guitar tabs (GTR-CTRL) based on desired instrumentation (inst-CTRL) and genre (genre-CTRL). Special control tokens are appended at the beginning of each song in the training corpus. We assess the performance of the model with and without conditioning. We propose instrument presence metrics to assess the inst-CTRL model's response to a given instrumentation prompt. We trained a BERT model for downstream genre classification and used it to assess the results obtained with the genre-CTRL model. Statistical analyses evidence significant differences between the conditioned and unconditioned models. Overall, results indicate that the GTR-CTRL methods provide more flexibility and control for guitar-focused symbolic music generation than an unconditioned model.

cs.SD↗

arXiv2022

Role of Cybersecurity and Blockchain in Battlefield of Things

The Internet of Things is an essential component in the growth of an ecosystem that enables quick and precise judgments to be made for communication on the battleground. The usage of the battlefield of things (BoT) is, however, subject to several restrictions for a variety of reasons. There is a potential for instances of replay, data manipulation, breaches of privacy, and other similar occurrences. As a direct result of this, the implementation of a security mechanism to protect the communication that occurs within BoT has turned into an absolute requirement. To this aim, we propose a blockchain-based solution that is both safe and private for use in communications inside the BoT ecosystem. In addition, research is conducted on the benefits of integrating blockchain technology and cybersecurity into BoT application implementations. This work elaborates on the importance of integrating cybersecurity and blockchain-based tools, techniques and methodologies for BoT.

cs.CR↗