7,731 papers in this slice of arXiv.
Zheng Xing, Chan-Tong Lam, Xiaochen Yuan
Classical image analysis routinely discards structurally meaningful orientation signatures encoded within Fourier phase, which are easily corrupted by local cellular rotation. Although quantum-inspired data processing offers new avenues for complex signal characterization, practical tools for directly extracting gauge-invariant angular correlations without explicit phase reconstruction remain scarce. Here we introduce Quantum Phase Bicoherence (QPBC) spectroscopy, a novel quantum-interferometric framework for capturing gauge-invariant angular order. The method embeds image angular sectors into a nine-qubit entangled state and probes three-body bicoherence via an ancilla, yielding 16 interpretable readout channels. We validate our framework on three independent public multi-modal imaging datasets covering fluorescence (BBBC021), bright-field (BBBC041) and histopathology (PathMNIST). QPBC consistently resolves angular-phase order and discriminates distinct biological phenotypes with high statistical significance. After principal-axis alignment, the optimal probing frequency universally converges, driven by Fourier directional sensitivity; negative-control experiments fully eliminate discriminative capacity, demonstrating frequency tuning acts as an on-off switch. Cross-dataset benchmarks confirm QPBC outperforms conventional Fourier-phase statistics, where inherent inversion symmetry serves as a built-in pipeline self-check. QPBC delivers a universal, classically unachievable quantitative texture observable, establishes interpretable quantum morphometry, and broadens the toolbox for quantum-inspired analysis applicable to diverse multi-modal microscopic measurements.
Clément Choné, Leslie Xu, Filippo Quadri +5
The growing deployment of real-time applications on wearable and Internet of Things (IoT) edge devices has intensified the need for energy-efficient, high-performance systems that meet stringent timing and energy constraints. Events-driven architectures leverage the sparsity of real-time to further improve system energy efficiency by employing an always-on (AO) domain to monitor inputs and activate a high-performance (HP) domain only when relevant events occur. However, for low-duty-cycle applications, the energy bottleneck shifts toward the AO domain, where leakage power dominates overall consumption. To mitigate this issue, AO circuits are typically implemented using high voltage threshold (HVT) or ultra-high voltage threshold (UHVT) transistors, thereby avoiding sub- and near-threshold operation, which is highly sensitive to process, voltage, and temperature (PVT) variations. In this context, feedforward leakage suppression logic (FLSL) has recently emerged as a promising candidate, offering reduced leakage compared to conventional. However, previous studies report a significant degradation in FLSL leakage performance in technology nodes below 90 nm, primarily due to increased gate and junction leakage currents. FDSOI technology, with its ability to effectively suppress junction leakage, provides an opportunity to overcome this limitation and restore FLSL efficiency in advanced nodes. Therefore, we demonstrate in this work that FLSL implemented in a 22 nm FDSOI technology can significantly reduce the energy consumption of small AO circuits with low-frequency inputs compared to state-of-the-art ultra-low-power CMOS designs. Silicon measurements on an FIR filter and an AES cryptographic core show reduced operating voltage and up to 9.8 x and 1.83 x reductions in leakage power compared to equivalent HVT and UHVT CMOS implementations, respectively.
Siddarth Gottumukkula, M P Samartha, Vedant Pahariya +3
This paper presents ProbSplat, a Compute-in-Memory (CIM)-inspired architecture based on programmable and energy efficient floating-gate inverter columns for probabilistic computing. Improving upon our prior work, ProbSplat programs and stores both means and variances of Gaussian mixture components, and evaluates log-likelihood for gaussian splatting during scene reconstruction with high energy efficiency, suitable for robotics and augmented/virtual reality (AR/VR) at the edge. Our proposed scheme enables independent control of both mean and variance via deterministic adjustment of floating-gate MOSFET threshold voltages, increasing the fidelity of hardware to program probability distributions. The design is simulated in 180nm CMOS on 1.8 V at 50 MHz and achieves mean-variance independence with <2.4% deviation during 3-D Gaussian mixture modeling. Compared to conventional digital implementations, ProbSplat significantly reduces compute complexity, memory footprint, and power consumption. The scalable framework consumes 18pJ energy per log-likelihood inference with 4-bit precision while operating for 500 mixture functions in a 3-D GMM. Scene reconstruction with ProbSplat's characteristics gave satisfactory fidelity of 21.99 PSNR (dB) at 8-bit precision.
Sanjeev Manivannan
Organizational decisions are co-created while evidence, constraints, and human priorities continue to evolve. In conventional transcript-based multi-agent systems, humans typically provide an initial problem, agents deliberate internally, and the system returns a final response. BoardroomAI instead treats the human as a persistent participant who can intervene by challenging assumptions, modifying constraints, changing priorities, introducing evidence, or redirecting the decision process. We operationalize this human--agent coexistence through four components: (i) a typed decision graph representing evidence, assumptions, constraints, claims, objections, alternatives, risks, decisions, semantic dependencies, and specialist responsibility; (ii) an intervention compiler that converts confirmed human actions into explicit graph updates; (iii) dependency-aware propagation that identifies affected subgraphs, preserves unaffected artifacts, and selectively reactivates relevant specialists; and (iv) an evaluation framework measuring intervention impact, repair coverage, preservation, recomputation, and decision validity. Across 600 generated decision-DAG interventions, propagation matched exhaustive impact computation while inspecting only 14.59% of nodes. In a 12-case exploratory pilot, selective repair recomputed 62.11% of canonical nodes, preserved all gold-unaffected nodes, and produced valid updated decisions in six cases while abstaining in the remaining six. These abstentions show that correct intervention routing may still provide insufficient context for synthesis, motivating a decision-sufficient context closure for human-steered multi-agent deliberation. All results are synthetic and prototype-level.
Qigao Zhu, Haojia Jiang, Guan Wang +4
Binarized Fourier single-pixel imaging (FSI) takes full advantage of the high modulation speed of digital micromirror devices by applying Floyd-Steinberg spatial dithering to binarize grayscale Fourier patterns. However, the use of the spatial dithering introduces substantial quantization errors, leading to decreasing imaging quality. Here, we propose a binarization method for grayscale Fourier patterns based on exponential multi-graylevel computational-weighted dithering, aimed at reducing quantization errors and then enhancing the imaging quality of binarized FSI. The proposed method quantizes Fourier patterns into 2R values {0,1/(R−1),2/(R−1)...1} and then decomposes them into binarized patterns. Both simulation and experimental results demonstrate that the method significantly reduces quantization errors in Fourier coefficients acquisition and improves imaging quality. The mean absolute percentage error of Fourier coefficients decreases from 194% to 28% and the structural similarity of reconstructed images (256×256 pixels) improves from 0.430 to 0.971, a 126% enhancement compared to the conventional method. Lateral resolution of this proposed method almost approaches the theoretical lateral resolution limit calculated by Rayleigh Criterion.
Harshkumar Oza, Aritra Sarkar, Syed Naqi Abbas +4
As quantum computing progresses from proof-of-principle demonstrations toward practical utility, a significant impediment is the need to augment algorithmic feasibility with system-level optimization across heterogeneous hardware and software stacks. Quantum resource estimation (QRE) plays a central role in this transition, yet existing approaches remain largely compilation-heavy or domain-knowledge-guided symbolic annotations, and tightly coupled to long-term fault-tolerant assumptions, limiting their topical applicability. In this work, we introduce AutoQuREO, an Automated framework for full-stack Quantum Resource Estimation and Optimization. AutoQuREO is built around four core novelties: (i) a flexible, user-defined abstraction of the quantum computing stack; (ii) a modular library of reusable stack components enabling rapid full-stack prototyping; (iii) surrogate modeling of layer-wise resources via algorithmic profiling and neuro-symbolic learning; and (iv) integrated multi-objective optimization that embeds QRE directly into deployment pipelines. Together, these design choices enable AutoQuREO to serve as a digital twin for quantum computing stacks, supporting the tractable exploration of complex design spaces. We demonstrate the capabilities of AutoQuREO through representative co-design case studies, including early-fault-tolerant quantum algorithms, small error correction codes, gate decomposition and variational training of parametric quantum circuits. These examples illustrate how AutoQuREO enables systematic discovery of unexploited resource trade-offs that are computationally intractable or abstruse using existing QRE tools. AutoQuREO is positioned as a general-purpose platform for advancing quantum technology readiness.
Yuanqiu Tan, Rahul Tripathi, Saleh Bunaiyan +5
Probabilistic bits (p-bits) have emerged as a cornerstone of probabilistic computing, enabling energy-efficient hardware implementation for probabilistic inference and combinatorial optimization. A critical challenge in advancing this field beyond binary p-bits lies in realizing and manipulating vector spin information, essential for mapping complex energy-based models such as the Heisenberg Hamiltonian.Here, we demonstrate a spintronic platform capable of real-space vector summation by using dual ferromagnetic spin injections into a monolayer graphene channel. By electrically tuning the spin polarization through independently controlled injection currents, we achieve continuous control over the magnitude and direction of the resulting spin accumulation vector. Experimental observations, supported by theoretical vector summation models and spin-circuit simulations, reveal coherent vector interactions and angular tunability of the spin state. This approach enables direct implementation of vector-based spin logic and lays the groundwork for mapping classical Heisenberg models using stochastic low-barrier magnets. Our results establish a scalable pathway for realizing probabilistic spin circuits based on two-dimensional materials, offering new opportunities for low-power, non-Boolean computing architectures.
Preety Priya, Yi Hong, Emanuele Viterbo
In this paper, we consider an orthogonal time frequency space (OTFS) system in time-varying channels with overspread Doppler shifts, typically found in non-terrestrial multi-satellite links. The overspread Doppler shifts with magnitude greater than half of the subcarrier spacing, result in aliased Doppler shifts in the delay-Doppler (DD) domain due to the OTFS modulo operation. This makes channel estimation very challenging and the traditional channel estimation methods become ineffective. To address this challenge, we propose a DD training frame and a two-stage channel estimation method. The training frame comprises a cosine pilot signal and a pilot symbol. In the first stage of the channel estimation, the pilot symbol in the DD domain is utilized to estimate the delays, aliased Doppler shifts, and channel gains of the propagation paths. In the second stage, the received time domain signal is converted into the frequency domain to detect the peaks of all the Doppler shifts using the cosine pilot signal. Then, we present a threshold-based method to pair the estimated actual Doppler shifts with their corresponding delays and channel gains. The complexity of the proposed channel estimation is also discussed. Finally, the performance of the proposed channel estimation is validated in terms of the normalized mean square error (NMSE) and bit error rate (BER) in various scenarios.
Nivedita Dey, Mrityunjay Ghosh, Pranav Kaushal +2
Efficient integration of cryptanalytic evidence with quantum verification remains a fundamental challenge in hybrid classical-quantum cryptanalysis. This work presents a capability-adaptive cryptanalytic framework that unifies linear cryptanalysis, differential cryptanalysis, and side-channel leakage analysis within a common candidate-space reduction architecture, followed by reduced-space quantum verification through amplitude amplification. A formal mathematical model is developed for candidate-space construction, adaptive filtering, verification-space reduction, and complexity characterization, supported by the oretical results establishing the relationship between candidate-space contraction and quantum verification effort. A Hamiltonian formulation is further introduced to provide a physically realizable interpretation of the reduced-space verification process. Evaluation using statistically generated cryptanalytic observations demonstrates that the proposed framework reduces an initial candidate-key hypothesis space of 4096 candidates to an effective candidate space of 13 hypotheses, corresponding to an overall reduction of approximately 99.683%. Consequently, the Grover verification requirement decreases from 50 iterations to only 2 iterations, yielding an approximately 25-fold reduction in verification effort, while reduced-space amplitude amplification achieves a target-state success probability of approximately 94.53%. These results demonstrate that adaptive cryptanalytic filtering can substantially reduce quantum verification complexity while preserving cryptanalytic admissibility, providing a practical foundation for capability-aware hybrid cryptanalysis and reduced-space quantum search.
Cyrill Bösch, Yigithan Gediz, Hakan Türeci
Physical computing systems exploit device dynamics for computation, but their gradient-based optimization is challenging: backpropagation through a digital twin suffers from model-reality gap. On-device gradient computation could resolve this issue, and a handful of theoretical and experimental studies have proposed ways to achieve it. Yet a unifying theory identifying when a physical system can compute the gradient of its own performance has been missing. Here we develop such a unification, based on the adjoint method: we identify sufficient conditions under which the adjoint field required for formally exact gradients can be generated on the same hardware that performs the computation. Linear and nonlinear systems obey fundamentally different conditions: for linear systems damping or gain is admissible provided reciprocity is preserved. For nonlinear trajectory systems the sufficient conditions are reciprocity of the linearized system and the existence of a time-reversal mirror. Algorithmically, the nonlinear case requires infinitesimal nudging, whereas linear systems admit a finite-amplitude experiment. We recover Equilibrium Propagation, Hamiltonian echo backpropagation, fully forward mode training and in situ gradient methods in integrated-photonic and free-space-optical systems. We further show that reciprocity is only the simplest instance of a more general intertwining condition, which extends exact on-device gradient computation to a class of non-Hermitian, non-reciprocal systems. Further generalizations include time-dependent parameters, Onsager-reciprocal dynamics and nonlinear, PT-symmetric Schrödinger equations. Our work provides a unified theoretical basis for formally exact physical learning algorithms and a template for constructing them across a range of physical systems.
Aditya Humnabadkar, Huaizhong Zhang, Ardhendu Behera
Natural-language-based scenario generation offers an intuitive means of describing rare and complex driving interactions, yet it is still uncertain whether training with language-structured data leads to truly adaptive control policies. We propose Language-Structured Relational Q-Learning, instantiated through an Ego-Centric Relational Q-Network (ERQ-Net), which jointly learns inter-vehicle relevance and action values from dynamic traffic graphs. Language descriptions define surrounding-vehicle behaviours during training, while prompts and semantic actor roles are hidden from the policy. ERQ-Net must therefore infer threat relevance solely from observable kinematics and interactions. Across 2,500 safety-critical scenarios, language-structured training improves test success from 49-52% to 55-58% and increases adversary-focused attention from 1.2x to 2.1x, demonstrating emergent threat awareness. However, this representational gain does not consistently translate into adaptive control: trained policies perform similarly to the best constant action, while a portfolio of simple policies solves 76% of scenarios. We formalise this discrepancy as a recognition-control gap and show that reward reweighting and margin shaping do not eliminate the resulting policy collapse. Evaluations of realism, criticality, semantic accuracy, and transfer of state-interface representations to CARLA further highlight both the strengths and the constraints of language-structured relational policy learning in safety-critical driving scenarios.
Tanya Shreedhar, Zuji Zhou, Nitinder Mohan +1
Multipath transports place a client's WiFi, cellular, and satellite networks under one connection, yet real-time video gains little from them. The scheduler that assigns packets to paths sees only bytes, so it cannot tell a keyframe that anchors a second of video from an enhancement frame whose loss costs one image. We show that the limiting factor is not a shortage of path diversity but the absence of a channel through which the application can name what the transport cannot see. Media over QUIC Transport (MoQT), already deployed on production CDNs, carries media as named Objects, and relays forward each Object's metadata without interpreting it. The knowledge the scheduler lacks therefore already flows through the subscriber's relay. We present MoMQ, a MoQT extension that turns this metadata into path decisions. Applications and relay operators install declarative rules at the edge relay that match Object metadata and express delivery preferences against labeled paths. The relay evaluates the rules mechanically, so it acts on video semantics while containing no video logic. On a live testbed spanning Starlink and WiFi, four rules cut P99.9 frame completion time from 384.7 ms under the best transport-only scheduler to 114.1 ms and reduce the required playback buffer by 61.5%. MoMQ is the only configuration that meets the 150 ms interactive latency target. Across relays in two countries and subscribers on two continents, the same rules apply unchanged and retain their advantage wherever the paths remain disjoint.
Masoud Shokrnezhad, Tarik Taleb
The Organic 6G vision of a network of networks spanning an edge-cloud continuum complemented by non-terrestrial resources requires, to realize its promise, service provisioning that is simple to operate, scalable across independently administered domains, and agile under domain churn (i.e., domains dynamically joining and leaving). Despite advances in cross-domain orchestration, many proposals rely on heavy integration fabrics, multi-layer coordinators, and deep telemetry pipelines that hinder deployability and amplify coordination overhead. We propose a lightweight, decentralized conversational orchestration framework based on Large Language Model (LLM)-driven domain agents. Each domain remains autonomous: an agent observes local state via tools, reasons in a closed loop, and exchanges summaries with neighboring agents over an Agent-to-Agent (A2A) overlay aligned with data-plane coupling. Fast feasible placement is enabled by periodic, routing-like dissemination of reachability advertisements (latency, bottleneck bandwidth, and compute capacity), while safe re-optimization, scaling, and migration are handled through event-driven requests and negotiation. To meet real-time constraints, we deploy a compact reasoning model trained with verifier-based self-verification and periodically refined online via shadow updates. Simulations show manageable, near-linear control-plane overhead as domains scale and during domain joins, and robust decision quality, including recovery after objective changes. We close by outlining future research directions for principled, secure, and uncertainty-aware agentic orchestration in Organic 6G.
George Yamoah Afrifa, Theophilus Ansah-Narh, Marcellin Atemkeng
Groundwater variability in Ghana remains poorly characterized due to limited long-term in-situ observations. This study investigates groundwater storage anomalies using GRACE-derived data from 2004-2024 combined with statistical analysis and unsupervised machine learning. Groundwater anomalies were standardized using Z-scores, while an ensemble-based Isolation Forest framework was applied for anomaly detection. The results revealed substantial temporal variability, with persistent groundwater deficits during 2004-2009 followed by increasing positive anomalies after 2018. A total of 12 anomalous months were identified, comprising 5 deficit and 7 surplus events, with the strongest anomalies associated with groundwater deficits. Spatial analysis showed more frequent deficit anomalies in northern Ghana and stronger surplus occurrence in southern regions. Comparison with statistical thresholds further indicated that the machine learning framework captured additional subtle deviations beyond conventional threshold-based methods. Overall, the integration of GRACE observations with unsupervised anomaly detection provides a practical framework for groundwater monitoring in data-scarce environments.
Aleksandar Anžel, Chisom Anyabolu, Leon Wimbes +12
Background: Deoxyribonucleic acid (DNA) data storage is a paradigm with great potential for ultra-dense and durable information preservation. However, the rapid proliferation of coding schemes, or codecs, each with their own design constraints and reporting practices, has led to a fragmented landscape that lacks a standardized comparative assessment. Methods: We developed an open-source, modular benchmarking platform that systematically integrates and evaluates state-of-the-art DNA storage encoding and decoding methods (codecs). Our approach uses a curated, diverse set of baseline data and applies multidimensional assessment criteria that are aligned with the consensus standard of the DNA Data Storage Alliance. These criteria include encoding/decoding throughput, computational efficiency, error correction performance across substitutions, insertions, and deletions, and cost efficiency. Results: The developed platform integrates standardized wrapper functions for encoding and decoding, allows for the integration of new methods, and automates reproducible evaluations with comprehensive visual and tabular reporting. Benchmarking both contemporary and classical codecs using their default parameters and multiple metrics demonstrates that no single algorithm is optimal across all evaluated dimensions. The trade-offs between information density, success rate, runtime, and cost are quantified and shown to be critical factors in the design of future-proof formats. Conclusions: Our work establishes a rigorously standardized, open-source evaluation framework that enables reproducible benchmarking, supports evidence-based codec selection, and provides the necessary foundation for translating DNA data storage from experimental research into deployable archival systems.
Chao Lu, Muralikrishnan Gopalakrishnan Meena, Eduardo Antonio Coello Perez +2
Fluid dynamics workloads are dominated by repeated solves of large, structured linear systems, motivating the search for quantum acceleration. The Variational Quantum Linear Solver (VQLS) is a leading near-term candidate, but practical deployment on hybrid quantum--high--performance computing (HPC) systems faces three persistent challenges: (i) the linear-combination-of-unitaries (LCU) encoding of the system matrix explodes in memory and runtime as the problem size grows, (ii) ansatz selection is largely empirical, with no clear link between standard circuit metrics and solver convergence, and (iii) end-to-end VQLS pipelines have rarely been exercised on production HPC hardware at non-trivial qubit counts. This work addresses these challenges through three contributions. First, we benchmark four matrix-encoding strategies---naive LCU, PennyLane-integrated, Fast Walsh--Hadamard Transform (FWHT)-based parallel Pauli decomposition, and an singular value decomposition (SVD)-based two-term LCU---and show that the FWHT approach reduces peak memory by up to 1298× on an 11×11 Hele--Shaw grid, while the SVD-based coherent VQLS delivers over 10,000× per-iteration speedup over standard Pauli-based VQLS at 8 qubits. Second, we evaluate 11 ansatz families with gradient-free and gradient-based optimizers on canonical Hele--Shaw flow, and find that expressibility and entanglement metrics correlate only weakly with VQLS convergence, motivating problem-aware ansatz design. Third, we deploy the full workflow on the OLCF Frontier supercomputer and successfully simulate a 15-qubit tridiagonal Toeplitz system on a single node. Together, these results establish a practical baseline for VQLS in hybrid quantum--HPC computation fluid dynamic (CFD) workflows and identify the remaining bottlenecks for larger problems.
R. Yagan, C. Swindells, I. T. Vidamour +5
Physical reservoir computing utilizes the intrinsic nonlinear and history-dependent dynamics of physical systems to perform machine-learning tasks with minimal training overhead. Here, we introduce a nanomagnetic reservoir computer based on a heterogeneous array of interconnected magnetic nanorings, combined with multi-channel planar Hall effect readout. The device comprises subarrays of rings with systematically varied track widths ranging from 500 nm to 300 nm, enabling access to the heterogeneous dynamics of geometrically diverse magnetic systems within a single reservoir. By applying time-varying input signals as modulations of a driving rotating magnetic field, we evaluate the nanoring reservoir's performance on nonlinear signal transformation and Mackey-Glass time-series prediction tasks. We find that combining outputs from multiple width-dependent channels significantly reduces the normalized root-mean-square error compared to single-channel readout, with the optimal channel combinations depending on task requirements. These results demonstrate that geometric heterogeneity provides an additional, experimentally accessible degree of freedom and complementary computational features. Principal component analysis further reveals that a reduced subset of correlated features captures most of the computationally relevant information while suppressing noise contributions. These results demonstrate that controlled geometric heterogeneity enhances reservoir expressivity and suggest a route toward scalable magnetic computing architectures in which multi-output magnetic metamaterials serve as configurable dynamical building blocks for device networks.
Tony Newton, Dan-Adrian German
Quantum mechanics is difficult to teach because its conceptual content and mathematical notation usually arrive together. Rudolph's misty-state language was designed to decouple those burdens; basis states are visual objects, clouds represent superposition, gates act by elementary replacement rules, and destructive interference appears as cancellation rather than as matrix calculation. An elementary ``misty-state'' language can make quantum circuits accessible to students before they master complex linear algebra. Development presented here was initiated/led by the first author. The contribution is not a replacement for complete graphical calculi such as ZX or sum-over-paths. It is a source-specific bridge from an intuitive educational notation to a mathematically explicit, executable, and falsifiable semantics.
Shiva Ahir
Translating technical requirements across languages can introduce semantic drift, altering numerical constraints, polarities, modalities, or other specification-critical meaning. IDRAAK is presented as an interpretable framework for detecting such drift using a language-independent Semantic Requirement Representation (SRR), with six detection workflows evaluated, ranging from deterministic comparison to multi-agent verification and few-shot prompting. On 890 synthetic perturbations across 300 requirements from 10 engineering domains, a single LLM call with six few-shot examples achieves MCC=0.888 and F1=0.983, outperforming the evaluated structured and multi-stage alternatives. Further evaluation on PAWS-X (805 pairs, 5 languages) and XNLI (700 pairs, 7 languages) exposes complementary strengths and limitations of structured and LLM-based approaches. Deterministic SRR comparison performs strongly on technical requirements (F1=0.898) but poorly on general-domain text (F1=0.012), while structured evidence improves performance on adversarial paraphrases. Post-hoc Platt scaling further improves confidence calibration. The results demonstrate that increased agentic complexity does not necessarily improve semantic-drift detection and that simple few-shot prompting can provide a strong and efficient alternative.
Sourav Das, Tanmay Joshi, Kripabandhu Ghosh
The carbon footprint of any deployed Large Language Model (LLM) accumulates during inference, where repeated use of the model substantially exceeds the one-time cost of fine-tuning. Yet most efficiency interventions target either pre-training scale or post-hoc compression. We ask whether folding a calibrated, differentiable energy surrogate into the fine-tuning objective can produce inference behavior that gains task accuracy at zero or near-zero carbon cost, a break-even configuration. We propose a joint loss mechanism with a per-model carbon-emission parameter, a linear surrogate over parameter norm, FLOP proxy, and a memory proxy, fit from on-hardware energy profiling. We fine-tune three architecturally distinct families: Gemma-2 2B, Llama-3.1 8B, and Qwen-2.5 14B, and evaluate inference F1 and CO2 emissions on three MMLU subjects: abstract algebra, philosophy, and formal logic. We discover from several outcomes that the carbon term behaves as either harmful interference or beneficial regularization depending on the task structure. We position calibrated carbon-aware fine-tuning as a lightweight, drop-in regularizer with a non-empty but model and task-dependent break-even region. This is an ongoing work, and we will release our codebase soon.