Search arXiv⌕ Search

arXiv subjects

Anirban Das

Publications and source records attributed to Anirban Das.

At least 19 recordsLinked to original sources
arXiv2026

Anisotropic g-tensor and strong spin-orbit interaction of holes in hexagonal germanium nanowires

Hexagonal germanium combines direct-gap optical activity with the low-nuclear-spin environment of a group-IV material, making it a candidate platform for spin qubits and spin-photon interfaces. Using a ten-band $\mathbf{k}\cdot\mathbf{p}$ model, we study the response of confined hole states in hexagonal germanium nanowires to magnetic and electric fields. We show that, by symmetry, the top valence band of the bulk crystal has a vanishing $g$ factor in the plane perpendicular to the $c$ axis, in analogy with the in-plane $g$ factor of heavy holes in two-dimensional hole gases. In nanowires, confinement induces only small transverse $g$-tensor components, so the $g$ tensor remains strongly anisotropic. A transverse electric field strongly reduces the nanowire band gap, providing a gate-tunable optical signature and limiting the usable field range. The same field induces a Rashba spin-orbit interaction that is markedly stronger for fields perpendicular to the $c$ axis and depends nonlinearly on the field strength. In narrow wires at strong fields, the spin-orbit length reaches a few tens of nanometers, comparable to hole devices in cubic silicon. We identify the microscopic origin of the Rashba coupling as a second-order process mediated by a few intermediate subbands with distinct symmetry selection rules. Our results provide guidance for the design of electrically controlled hole-spin qubits in hexagonal germanium nanowires.

cond-mat.mes-hall↗
arXiv2026

Latest $ν$s from the Earliest Galaxies: Probing neutrino self-interactions with JWST

Our Universe serves as a complementary laboratory for particle physics interactions. One intriguing possibility is that neutrinos may have beyond-the-Standard-Model (BSM) self-interactions, which can delay their free streaming and have an observable effect across different redshifts. Such interactions can induce scale-dependent changes in the matter power spectrum. The James Webb Space Telescope (JWST) has opened up a new window onto the early Universe by discovering a large population of high-redshift galaxies up to $z \sim 15$ and various probable galaxy candidates at even higher redshifts, between $z \sim 15-30$. Such high-redshift observations can probe the matter power spectrum on small scales, since structure formation is hierarchical and the structures hosting these high-redshift galaxies are therefore expected to be small. We, $\textit{for the first time}$, use the UV luminosity function data from JWST and combine them with the $\textit{Planck}$ CMB data to probe neutrino self-interactions ($ν$SI). If only one neutrino flavor experiences such BSM interactions, then $\textit{Planck}$ + JWST data give a 95% C.L. upper limit on $\mathrm{Log}_{10}(G_{\mathrm{eff}}/\mathrm{MeV}^{-2})$ $\approx$ $-2.75$ and when all three flavors are interacting, the corresponding constraint becomes $\lesssim -3.41$; both stronger than those obtained using only $\textit{Planck}$ data. Our results highlight the power of this new observable, the high-redshift UV luminosity function from JWST as a complementary probe to terrestrial experiments, showing that small-scale observables can open a new window onto smaller neutrino self-interaction strengths.

astro-ph.CO↗
arXiv2026

Probing freeze-in dark matter using Bose-Einstein condensate in neutron star

Neutron star (NS) is one of the most promising astrophysical targets to probe non-gravitational interaction of dark matter (DM) with visible matter. Their compactness makes them an ideal object which can capture particle DM efficiently over its lifetime using the DM-nucleon scattering cross-section. If DM particles are bosonic, then the captured DM population may form a Bose-Einstein condensate at the center of the NS, increasing the DM density significantly. In this work, we study the phenomenology of such scenario with enhanced DM annihilation rate due to the increased density in a condensate. The enhanced DM annihilation makes the NS surface `hotter' than in the standard cooling scenario. We show that the annihilation rate is enhanced by a factor of $\mathcal{O}(10^{15}-10^{20})$ if DM forms a condensate without large self-interaction, and DM with freeze-in value annihilation cross-section can heat up the NS to higher temperatures, bringing it within the reach of James Webb Space Telescope. It also allows us to probe DM-nucleon scattering cross section within the neutrino fog regime which will complement the terrestrial direct detection searches. Moreover, the enhanced annihilation from the condensate changes the lower limits on s-wave DM annihilation cross-section for capture-annihilation equilibrium and the formation of a black hole inside the NS. Finally, we show an example of a scalar DM model where such small annihilation and DM-nucleon scattering cross sections can generically arise.

hep-ph↗
arXiv2026

Engineering Ergodicity-Breaking and Stable Emergent Conservation Laws in a Generic Quantum Many-body System Under Strong Periodic Drive

A periodically driven quantum many-body system is believed to approach a locally infinite-temperature-like state as dictated by the ergodicity hypothesis, when all the couplings/parameters are of the same order. When the drive is stronger than a threshold value, a large class of systems is shown to break this mandate, exhibiting Dynamical Freezing (DF): ergodicity is broken due to the emergence of local operators with permanent memory of their initial values (PRX 11, 021008, 2021). However, an exception to DF has recently been noted in a particular setting (PRX 16, 011041, 2026). Here we identify the condition for ergodicity-breaking in a system under strong drive, and show how to stabilize DF by exact, local engineering. We further show that the Floquet Hamiltonian $H_{eff} = (i/T)\ln{[U(T)]}$ is non-local when ordered by spin-flips, in a non-DF system, while it turns quasi-local under our simple, local DF engineering that keeps the Hamiltonian $H(t)$ still non-integrable at each $t$. This perplexing non-universality is absent in equilibrium, where non-integrability is believed to imply ergodicity and thermalization universally. Our work paves a way to engineering novel ergodicity-broken interacting Floquet phases of matter in various quantum platforms, and also endows DF-assisted quantum error mitigation and state stabilization techniques (see Nature 653, 1027 (2026)) with unprecedented versatility and flexibility.

quant-ph↗
arXiv2026

GAIA meets LZ: a high velocity tail-tale sign for dark matter

The interpretation of high-energy nuclear recoils in direct-detection experiments can depend sensitively on the high-velocity tail of the galactic dark matter (DM) distribution. The LUX-ZEPLIN (LZ) experiment has recently reported an event, LZ230616, with nuclear recoil energy 248 $\pm$ 23 (stat) $\pm$ 23 (sys) keV, while the Gaia mission has provided an unprecedented view of the kinematics and mass distribution of the Milky Way. We bring these two probes together to investigate the inelastic DM interpretation of LZ230616, using a velocity distribution inferred from Gaia-informed galactic model. Compared to the Standard Halo Model, the resulting distribution has a lower escape speed and a significantly suppressed high-velocity tail, with a higher local density. We find that while the exothermic interpretation is largely insensitive to this change, the endothermic interpretation requires substantially larger scattering cross sections and smaller mass splittings. Furthermore, using simulation-based study of the gravitational effects of the Large Magellanic Cloud, we show that the high-velocity tail can extend the allowed endothermic parameter space towards lower DM masses and larger splittings. Our results demonstrate how combining direct searches for DM detection with precision measurements of galactic dynamics can qualitatively affect the particle-physics interpretation of high-energy nuclear recoils.

hep-ph↗
arXiv2026

Learning to Stop without Learning to Stop: Self-Supervised Confidence Training Improves Reasoning Efficiency

Reasoning models often generate very long reasoning traces, making inference computationally expensive. Existing approaches typically improve efficiency either through inference-time early-stopping mechanisms or by explicitly encouraging shorter reasoning during training, for example through reinforcement learning with length penalties. We show that substantial efficiency gains can instead emerge from a different kind of supervision: \textit{confidence}. Using a self-supervised procedure, we fine-tune reasoning models to predict their confidence in the answer at intermediate points along their own reasoning trajectories using only 600 training problems. Confidence is used only as a training target: the loss contains no objective for reasoning length, efficiency, or stopping. At inference, the fine-tuned models use the standard generation procedure, with no confidence elicitation or early-stopping mechanism. Despite this, self-supervised confidence fine-tuning makes reasoning more efficient, reducing generated tokens by up to 25\% at matched accuracy across Gemma, Qwen, Nemotron, and GPT-OSS models on mathematical, scientific, and coding reasoning benchmarks, with efficiency gains comparable to methods that explicitly optimize for shorter reasoning. Analysis of reasoning episodes further shows that confidence supervision largely preserves the base models' high-level reasoning composition rather than selectively suppressing particular behaviors. Our results suggest that efficient reasoning may emerge as a downstream consequence of learning metacognitive signals, without being directly optimized.

cs.AI↗
arXiv2026

Parabolic defect measures for Harmonic Maps with Free Boundary

We introduce defect measures for the heat flow of a class of harmonic maps with free boundary and initiate their analysis. Our flow is motivated by singular perturbations on the boundary of the domain, emphasizing structural differences between harmonic maps with free boundary and their classical counterpart. The harmonic maps we consider are a reformulation of maps introduced by Da Lio and Rivière and arise in several problems in the spectral geometry of Steklov eigenvalues. Contrary to the classical case, the flow we consider exhibits features of global nature and the blow-up analysis turns out to be much more challenging.

math.AP↗
arXiv2026

The Rise of Verbal Reinforcement Learning

Natural language is emerging as a primary feedback channel for improving language agents, capable of conveying intent, preferences, and causal structure in forms interpretable by both humans and modern language models. We call this paradigm Verbal Reinforcement Learning (VRL) and offer the first unified account of it. We organize the field around a single axis, \textit{when} verbal feedback takes effect in an agent's lifecycle and \textit{what} it modifies, yielding three pillars: (1) \textbf{Language as Grounding Signal}, where language defines the task itself by specifying goals, states, and reward structures; (2) \textbf{Language as Deliberative Feedback}, where natural language guides reasoning at test time without the need to update model parameters; (3) \textbf{Language as Learning Signal}, where language-based feedback shapes model parameters through training. Within each pillar, we synthesize representative work, distinguish key subcategories of approaches, and outline the distinct role language plays in shaping agent behavior. Together, this taxonomy shows how verbal reinforcement is reshaping agent development, while also defining the challenges and opportunities for building more capable and aligned agents.

cs.CL↗
arXiv2026

AcCoRD: Evaluating User-Agent Collaboration Under Realistic User Preference Dynamics

User preferences in user-agent collaboration are rarely static and fully-specified upfront: preferences are formed, revealed, adjusted, and relaxed during interaction. Existing benchmarks for evaluating user-agent collaboration focus almost exclusively on resolving underspecified preferences, thereby failing to capture the richer dynamics of real-world interaction. We introduce AcCoRD, a user-agent collaboration benchmark requiring agents to handle diverse user preference dynamics in two domains: online shopping and travel planning. We evaluate five frontier LLMs under two prompting strategies: vanilla ReAct, and an uncertainty-guided variant that prompts models to identify and resolve ambiguity about user preferences. Our results reveal that frontier models can handle underspecification but struggle to satisfy preferences that emerge or evolve mid-interaction and require more sophisticated uncertainty modeling. Further, prompting alone fails to elicit the required uncertainty recognition. We release AcCoRD as a resource for developing agents that can navigate the full complexity of real-world user preferences.

cs.AI↗
arXiv2026

Structured Thoughts For Improved Reasoning And Context Pruning

Large language models (LLMs) excel at generating long chains of thought, but long reasoning traces are often verbose and memory-inefficient. In this work, we introduce Structured Thoughts, a framework that organizes reasoning into alternating and blocks: captures exploratory scratch work, while contains the distilled conclusion of that step. We construct a dataset of structured thoughts by segmenting reasoning traces into blocks and prompting an LLM to summarize each step into its corresponding . Fine-tuning pretrained foundation models on this reformatted data produces models that adopt the structured reasoning style, leading to performance gains of up to 8.08\% on reasoning benchmarks compared to standard SFT. The explicit structure also enables context pruning: after each / pair, the can be pruned, allowing the model to retain conclusions without keeping the full scratch work in the context. A proof-of-concept pruning implementation achieves an average of 85\% memory / context savings with an 8.67\% performance drop across mathematical tasks.

cs.CL↗
arXiv2026

Project Auto-World: Towards Automated Benchmarking of Neural Relational Reasoners

Reasoning about relational structures remains a significant challenge for neural models, particularly when they must systematically apply learned knowledge to problem instances that are harder than those seen in training. Progress is hampered by the difficulty of evaluating such generalization, since a priori, it is rarely clear what makes an instance hard. We study how this issue can be addressed by using large language models (LLMs) to automate benchmark generation, learning to produce increasingly challenging instances in an end-to-end manner. Concretely, given a world parametrized by Datalog rules, and an Edge Transformer as the reasoning evaluator, we use LLM-driven evolutionary search (based on FunSearch) and autonomous agentic search to discover sampling functions that yield hard problem instances. We also show that the Edge Transformer can be improved using this data such that it generalizes well to further data perturbations. Finally, we show that the same machinery can be applied to novel worlds proposed by LLMs, opening the door to autonomous research on neural relational reasoning.

cs.AI↗
arXiv2026

Know Thy Reasoner: Not All Language Models Explore Alike

Compute scaling for LLM reasoning trades off exploring solution approaches (\emph{breadth}) against refining promising ones (\emph{depth}), yet why a given trade-off works, and why it often fails to transfer across models, remains unclear. We argue that \textbf{the optimal strategy depends on the model's \emph{diversity profile}, the spread of probability mass across solution approaches, and that this must be characterized before any exploration strategy is adopted.} We formalize this with a framework decomposing reasoning uncertainty, deriving when depth-based refinement outperforms parallel sampling, and validate it across three model families at both inference and training. Our central finding is that the diversity regime dictates the strategy: low-diversity aligned models benefit from depth-based refinement with lightweight intrinsic signals, whereas high-diversity base models are often harmed by it, and instead need breadth or stronger signals to compensate.

cs.AI↗
arXiv2026

T1-Bench: Benchmarking Multi-Scenario Agents in Real-World Domains

Recent advances in reasoning and tool-calling capabilities of large language models (LLMs) have enabled increasingly capable agentic systems. However, existing benchmarks remain limited in task complexity, realism, and domain diversity, and often fail to capture interactions that span multiple domains, limiting their ability to evaluate agents in realistic multi-step settings that require sustained reasoning and coordination. To address these limitations, we introduce T1-Bench, a high-fidelity, comprehensive benchmark for evaluating agentic systems in realistic customer-facing, multi-domain environments, featuring interleaved scenarios that require structured reasoning across multi-turn user-assistant interactions and substantially increasing both compositional complexity and evaluative rigor across 25 domains of varying difficulty. We evaluate T1-Bench using 12 proprietary and open-weight models, providing a reproducible and standardized framework for assessing agent behavior, tool utilization, and conversational quality in complex, multi-step environments. We further complement automatic evaluation with human judgments to strengthen the assessment of qualitative performance. Overall, T1-Bench substantially advances prior benchmarks by increasing task complexity, interaction depth, and domain coverage in simulated multi-domain environments. To facilitate future research on agentic systems, we will publicly release data and evaluation code as open source.

cs.CL↗
arXiv2026

AVSD: Adaptive-View Self-Distillation by Balancing Consensus and Teacher-Specific Privileged Signals

Self-distillation enables language models to learn on-policy from their own trajectories by using the same model as both student and teacher, with the teacher being conditioned on privileged information unavailable to the student. Such information can come in different types or views, such as solutions, demonstrations, feedback, or final answers. This setup provides dense token-level feedback without relying on a separate external model, but creates a fundamental asymmetry: the teacher may rely on view-specific information that the student cannot access at inference time. Moreover, the best type of privileged information is often task-dependent, making it difficult to choose a single teacher view. In this work, we address both these challenges jointly by introducing AVSD (Adaptive-View Self-Distillation), a novel method of self-distillation with multiple privileged-information views, which reconstructs token-level supervision by separating stable cross-view consensus from view-specific residual signals. AVSD identifies the consensus signal shared across views, which provides a reliable update direction, and then selectively adds the view-specific residual signal to adjust the update magnitude when it both aligns with the consensus direction and remains proportionate to the consensus signal. Experiments on math competition benchmarks (AIME24, AIME25, and HMMT25) show that AVSD consistently outperforms both single-view self-distillation baselines and GRPO, achieving average Avg@8 gains of 3.1% and 2.2% over the strongest baselines on Qwen3-8B and Qwen3-4B, respectively. Moreover, on code-generation benchmarks (Codeforces, LiveCodeBench v6) using Qwen3-8B, AVSD outperforms the single-view self-distillation baseline by 2.4% on average.

cs.LG↗
arXiv2026

Dive deeper with SUBMARINE: SUB-Mev dArk matter diRect detectIon using bilayer grapheNE

Novel target materials with anisotropic response will play a key role in detecting low-mass dark matter in upcoming experiments. Bilayer graphene is one such material that has been proposed for the detection of sub-MeV mass dark matter particles via electronic excitations. In this work, we calculate scattering rate via a massive mediator in bilayer graphene. With an exposure as small as $\sim$ 0.5 mg-year, bilayer graphene can probe new regions of the parameter space. The anisotropic response function of bilayer graphene leads to a sidereal-day modulation in the scattering rate, depending on its orientation with respect to the Galactic dark matter wind. We find significant modulation in the scattering rate for sub-MeV mass dark matter, demonstrating bilayer graphene's promise for a future experiment. We hope that our work will motivate the community to investigate bilayer graphene as a novel target material, and that it may lead us to discover the particle nature of dark matter.

hep-ph↗
arXiv2026

Impostor Among $ν$s: Dark Radiation Masquerading as Self-Interacting Neutrinos

Multiple cosmological observations hint at neutrino self-interactions beyond the Standard Model, yet such interactions face severe constraints from terrestrial experiments. We resolve this tension by introducing a model where active neutrinos resonantly convert to self-interacting dark radiation after BBN but before CMB epoch. This exploits the fact that cosmological observables cannot distinguish between neutrinos and dark radiation with the same abundance and free-streaming properties. Our mechanism, based on a simple type-I seesaw framework along with a keV-scale scalar mediator, achieves two objectives: (i) it produces strongly self-interacting dark radiation that imitates neutrino self-interactions favored by cosmological data, and (ii) it depletes the active neutrino energy density, relaxing cosmological neutrino mass bounds and easing the tension with neutrino oscillation data. The model naturally evades laboratory constraints through suppression of the neutrino-mediator coupling by the squared mass ratio of active and sterile neutrinos. We show that this scenario is favored over $Λ$CDM by the combined Planck and DESI data, while being consistent with all other constraints. Our mechanism is testable in future laboratory probes of absolute neutrino mass and searches for sterile neutrinos.

hep-ph↗
arXiv2026

Can Large Language Models Understand, Reason About, and Generate Code-Switched Text?

Code-switching is a pervasive phenomenon in multilingual communication, yet the robustness of large language models (LLMs) in mixed-language settings remains insufficiently understood. In this work, we present a comprehensive evaluation of LLM capabilities in understanding, reasoning over, and generating code-switched text. We introduce CodeMixQA a novel benchmark with high-quality human annotations, comprising 16 diverse parallel code-switched language-pair variants that span multiple geographic regions and code-switching patterns, and include both original scripts and their transliterated forms. Using this benchmark, we analyze the reasoning behavior of LLMs on code-switched question-answering tasks, shedding light on how models process and reason over mixed-language inputs. We further conduct a systematic evaluation of LLM-generated synthetic code-switched text, focusing on both naturalness and semantic fidelity, and uncover key limitations in current generation capabilities. Our findings reveal persistent challenges in both reasoning and generation under code-switching conditions and provide actionable insights for building more robust multilingual LLMs. We release the dataset and code as open source.

cs.CL↗
arXiv2025

Continual Pre-training of MoEs: How robust is your router?

Sparsely-activated Mixture of Experts (MoE) transformers are promising architectures for foundation models. Compared to dense transformers that require the same amount of floating-point operations (FLOPs) per forward pass, MoEs benefit from improved sample efficiency at training time and achieve much stronger performance. Many closed-source and open-source frontier language models have thus adopted an MoE architecture. Naturally, practitioners will want to extend the capabilities of these models with large amounts of newly collected data without completely re-training them. Prior work has shown that a simple combination of replay, learning rate re-warming, and re-decaying can enable the continual pre-training (CPT) of dense decoder-only transformers with minimal performance degradation compared to full re-training. In the case of decoder-only MoE transformers, however, it is unclear how the routing algorithm will impact continual pre-training performance: 1) do the MoE transformer's routers exacerbate forgetting relative to a dense model?; 2) do the routers maintain a balanced load on previous distributions after CPT?; 3) are the same strategies applied to dense models sufficient to continually pre-train MoE LLMs? In what follows, we conduct a large-scale study training a 500M parameter dense transformer and four 500M-active/2B-total parameter MoE transformers. Each model is trained for 600B tokens. Our results establish a surprising robustness to distribution shifts for MoEs using both Sinkhorn-Balanced and Z-and-Aux-loss-balanced routing algorithms, even in MoEs continually pre-trained without replay. Moreover, we show that MoE LLMs maintain their sample efficiency (relative to a FLOP-matched dense model) during CPT and that they can match the performance of a fully re-trained MoE at a fraction of the cost.

cs.LG↗