Search

Building relevance models to rank documents based on user information needs is a central task in information retrieval and the NLP community. Beyond the direct ad-hoc search setting, many knowledge-intense tasks are powered by a first-stage retrieval stage for context selection, followed by a more involved task-specific model. However, most first-stage ranking stages are inherently limited by the recall of the initial ranking documents. Recently, adaptive re-ranking techniques have been proposed to overcome this issue by continually selecting documents from the whole corpus, rather than only considering an initial pool of documents. However, so far these approaches have been limited to heuristic design choices, particularly in terms of the criteria for document selection. In this work, we propose a unifying view of the nascent area of adaptive retrieval by proposing, Quam, a \textit{query-affinity model} that exploits the relevance-aware document similarity graph to improve recall, especially for low re-ranking budgets. Our extensive experimental evidence shows that our proposed approach, Quam improves the recall performance by up to 26\% over the standard re-ranking baselines. Further, the query affinity modelling and relevance-aware document graph modules can be injected into any adaptive retrieval approach. The experimental results show the existing adaptive retrieval approach improves recall by up to 12\%. The code of our work is available at \url{https://github.com/Mandeep-Rathee/quam}., Comment: 15 pages, 10 figures

In this paper, we explore linear representations of skew left braces, which are known to provide bijective non-degenerate set-theoretical solutions to the Yang--Baxter equation that are not necessarily involutive. A skew left brace $(A, \cdot, \circ)$ induces an action $\lambda^{\op}: (A, \circ) \to \Aut (A, \cdot)$, which gives rise to the group $\Lambda_{A^{\op}} = (A, \cdot) \rtimes_{\lambda^{\op}} (A, \circ)$. We prove that if $A$ and $B$ are isoclinic skew left braces, then $\Lambda_{A^{\op}}$ and $\Lambda_{B^{\op}}$ are also isoclinic under some mild restrictions on the centers of the respective groups. Our key observation is that there is a one-to-one correspondence between the set of equivalence classes of irreducible representations of $(A, \cdot, \circ)$ and that of the group $\Lambda_{A^{\op}}$. We obtain a decomposition of the induced representation of the additive group $(A, \cdot)$ and of the multiplicative group $(A, \circ)$ corresponding to the regular representation of the group $\Lambda_{A^{\op}}$. As examples, we compute the dimensions of the irreducible representations for several skew left braces with prime power orders., Comment: 17 pages, added Proposition 3.3: Proof that if $A$ is any skew left brace, then $\Lambda_{A}$ and $\Lambda_{A^{\op}}$ are isomorphic groups

Many inference services based on large language models (LLMs) pose a privacy concern, either revealing user prompts to the service or the proprietary weights to the user. Secure inference offers a solution to this problem through secure multi-party computation (MPC), however, it is still impractical for modern LLM workload due to the large overhead imposed by MPC. To address this overhead, we propose Marill, a framework that adapts LLM fine-tuning to minimize MPC usage during secure inference. Marill introduces high-level architectural changes during fine-tuning that significantly reduce the number of expensive operations needed within MPC during inference, by removing some and relocating others outside MPC without compromising security. As a result, Marill-generated models are more efficient across all secure inference protocols and our approach complements MPC-friendly approximations for such operations. Compared to standard fine-tuning, Marill results in 3.6-11.3x better runtime and 2.4-6.9x better communication during secure inference across various MPC settings, while typically preserving over 90% performance across downstream tasks.

We develop a theoretical framework to describe how zero-photon detection may be utilized to enhance laser cooling via the anti-Stokes interaction and, somewhat surprisingly, enable cooling via the Stokes interaction commonly associated with heating. Our description includes both pulsed and continuous measurements as well as optical detection efficiency and open-system dynamics. For both cases, we discuss how the cooling depends on the system parameters such as detection efficiency and optomechanical cooperativity, and we study the continuous-measurement-induced dynamics, contrasting to single-photon detection events. For the Stokes case, we explore the interplay between cooling and heating via optomechanical parametric amplification, and we find the efficiency required to cool a mechanical oscillator via zero-photon detection. This work serves as a companion article to the recent experiment [E. A. Cryer-Jenkins, K. D. Major, et al., arXiv:2408.01734 (2024)], which demonstrated enhanced laser cooling of a mechanical oscillator via zero-photon detection on the anti-Stokes signal. The framework developed here provides new approaches for cooling mechanical resonators that can be applied to a wide range of areas including nonclassical state preparation, quantum thermodynamics, and avoiding the often unwanted heating effects of parametric amplification., Comment: 15 pages, 6 figures

Throughout quantum science and technology, measurement is used as a powerful resource for nonlinear operations and quantum state engineering. In particular, single-photon detection is commonly employed for quantum-information applications and tests of fundamental physics. By contrast, and perhaps counter-intuitively, measurement of the absence of photons also provides useful information, and offers significant potential for a wide range of new experimental directions. Here, we propose and experimentally demonstrate cooling of a mechanical resonator below its laser-cooled mechanical occupation via zero-photon detection on the anti-Stokes scattered optical field and verify this cooling through heterodyne measurements. Our measurements are well captured by a stochastic master equation and the techniques introduced here open new avenues for cooling, quantum thermodynamics, quantum state engineering, and quantum measurement and control., Comment: Main: 5 pages, 2 figures. Supplemental: 6 pages, 2 figures

Model-Based Reinforcement Learning (MBRL) has been widely studied for Heating, Ventilation, and Air Conditioning (HVAC) control in buildings. One of the critical challenges is the large amount of data required to effectively train neural networks for modeling building dynamics. This paper presents CLUE, an MBRL system for HVAC control in buildings. CLUE optimizes HVAC operations by integrating a Gaussian Process (GP) model to model building dynamics with uncertainty awareness. CLUE utilizes GP to predict state transitions as Gaussian distributions, effectively capturing prediction uncertainty and enhancing decision-making under sparse data conditions. Our approach employs a meta-kernel learning technique to efficiently set GP kernel hyperparameters using domain knowledge from diverse buildings. This drastically reduces the data requirements typically associated with GP models in HVAC applications. Additionally, CLUE incorporates these uncertainty estimates into a Model Predictive Path Integral (MPPI) algorithm, enabling the selection of safe, energy-efficient control actions. This uncertainty-aware control strategy evaluates and selects action trajectories based on their predicted impact on energy consumption and human comfort, optimizing operations even under uncertain conditions. Extensive simulations in a five-zone office building demonstrate that CLUE reduces the required training data from hundreds of days to just seven while maintaining robust control performance. It reduces comfort violations by an average of 12.07% compared to existing MBRL methods, without compromising on energy efficiency.

Spatio-temporal evolution of large amplitude upper hybrid oscillations in a cold homogeneous plasma in the presence of an inhomogeneous magnetic field is studied analytically and numerically using the Dawson sheet model. It is observed that the inhomogeneity in magnetic field which causes the upper hybrid frequency to acquire a spatial dependence, results in phase mixing and subsequent breaking of the upper hybrid oscillations at arbitrarily low amplitudes. This result is in sharp contrast to the usual upper hybrid oscillations in a homogeneous magnetic field where the oscillations break within a fraction of a period when the amplitude exceeds a certain critical value. Our perturbative calculations show that the phase mixing (wave breaking) time scales inversely with the amplitude of magnetic field inhomogeneity ($\Delta$) and amplitude of imposed density perturbation ($\delta$), and scales directly with the ratio of magnetic field inhomogeneity scale length to imposed density perturbation scale length ($(\alpha/k_L)^{-1}$ ) as $\omega_{pe}\tau_{mix} \sim \left( 1+\beta^2 \right) ^{3/2}k_L/(\beta^2\delta\Delta\alpha)$, where $\beta$ is the ratio of electron cyclotron frequency to electron plasma frequency. Further phase mixing time measured in simulations, performed using a 1-1/2 D code based on Dawson sheet model, shows good agreement with the above mentioned scaling. This result may be of relevance to plasma based particle acceleration experiments in the presence of a transverse inhomogeneous magnetic field.

Rota--Baxter operators over groups have been recently defined in \cite{LHY2021}, and they share a close connection with skew braces, as demonstrated in \cite{VV2022}. In this paper, we classify all Rota--Baxter operators of weight 1 over the Heisenberg Lie algebra of dimension 3 by directly solving the operators defining equations. Using the fact that the exponential map from the Heisenberg Lie algebra to the Heisenberg Group is bijective, we induces these operators to the Heisenberg Group. Finally, we enumerate all skew left brace structures over the Heisenberg Group induced by these Rota--Baxter operators., Comment: 14 pages. The mistake in the matrix multiplication given to compute skew left brace structures has now been corrected

Relative Rota--Baxter groups, a generalization of Rota--Baxter groups, are closely connected to skew left braces, which play a fundamental role in understanding non-degenerate set-theoretical solutions to the Yang-Baxter equation. In this paper, we explore the inducibility problem for automorphisms of abelian extensions of relative Rota--Baxter groups. This problem is intricately linked to the recently introduced second cohomology of relative Rota--Baxter groups. Specifically, we prove a Wells-like exact sequence for abelian extensions of relative Rota--Baxter groups. The sequence establishes a connection among the group of derivations, certain automorphism group, and the second cohomology of relative Rota--Baxter groups, thereby giving precise structural relationships between these groups., Comment: 22 pages. arXiv admin note: text overlap with arXiv:2309.00692, arXiv:2311.12384

This review explores machine unlearning (MUL) in recommendation systems, addressing adaptability, personalization, privacy, and bias challenges. Unlike traditional models, MUL dynamically adjusts system knowledge based on shifts in user preferences and ethical considerations. The paper critically examines MUL's basics, real-world applications, and challenges like algorithmic transparency. It sifts through literature, offering insights into how MUL could transform recommendations, discussing user trust, and suggesting paths for future research in responsible and user-focused artificial intelligence (AI). The document guides researchers through challenges involving the trade-off between personalization and privacy, encouraging contributions to meet practical demands for targeted data removal. Emphasizing MUL's role in secure and adaptive machine learning, the paper proposes ways to push its boundaries. The novelty of this paper lies in its exploration of the limitations of the methods, which highlights exciting prospects for advancing the field., Comment: In Proceedings of 7th INTERNATIONAL CONFERENCE ON INNOVATIVE COMPUTING AND COMMUNICATION 2024 (https://icicc-conf.com/)

Relative Rota-Baxter groups are generalizations of Rota-Baxter groups and share a close connection with skew left braces. These structures are well-known for offering bijective non-degenerate set-theoretical solutions to the Yang-Baxter equation. This paper builds upon the recently introduced extension theory and low-dimensional cohomology of relative Rota-Baxter groups. We prove an analogue of the Hochschild-Serre exact sequence for central extensions of relative Rota-Baxter groups. We introduce the Schur multiplier $M_{RRB}(\mathcal{A})$ of a relative Rota-Baxter group $\mathcal{A} =(A,B,\beta,T)$, and prove that the exponent of $M_{RRB}(\mathcal{A})$ divides $|A||B|$ when $\mathcal{A}$ is finite. We define weak isoclinism of relative Rota-Baxter groups, introduce their Schur covers, and prove that any two Schur covers of a finite bijective relative Rota-Baxter group are weakly isoclinic. The results align with recent results of Letourmy and Vendramin for skew left braces., Comment: 29 pages. arXiv admin note: text overlap with arXiv:2309.00692

Relative Rota-Baxter groups are generalisations of Rota-Baxter groups and recently shown to be intimately related to skew left braces, which are well-known to yield bijective non-degenerate solutions to the Yang-Baxter equation. In this paper, we develop an extension theory of relative Rota-Baxter groups and introduce their low dimensional cohomology groups, which are distinct from the ones known in the context of Rota-Baxter operators on Lie groups. We establish an explicit bijection between the set of equivalence classes of extensions of relative Rota-Baxter groups and their second cohomology. Further, we delve into the connections between this cohomology and the cohomology of associated skew left braces. We prove that for bijective relative Rota-Baxter groups, the two cohomologies are isomorphic in dimension two., Comment: 30 pages

Dynamics of double Jaynes-Cummings models are studied in the presence of Markovian noise and cavity disorders with specific attention to entanglement sudden death and revivals. The study is focused on the glassy disorders, which remain unchanged during the observations. The field is initially assumed to be in a vacuum state, while the atoms are considered to be in a specific two-qubit superposition state. Specifically, the study has revealed that the presence of noise, or a nonlinear pump results in interesting behaviors in the entanglement dynamics. Further, entanglement sudden death is observed in the presence of Markovian noise and nonlinear pump. The presence of entanglement sudden deaths and revivals have also been observed in cases where they were absent initially for the chosen states. The effect of noise on the dynamics of the system is to decay the characteristics, while that of the disorder is to wash them out. On the other hand, the introduction of nonlinearity is found to cause the dynamics of the system to speed up., Comment: Entanglement dynamics of variants of double Jaynes-Cummings models are studied

Relative Rota-Baxter groups are generalisations of Rota-Baxter groups and introduced recently in the context of Lie groups. In this paper, we explore connections of relative Rota-Baxter groups with skew left braces, which are well-known to give non-degenerate set-theoretic solutions of the Yang-Baxter equation. We prove that every relative Rota-Baxter group gives rise to a skew left brace, and conversely, every skew left brace arises from a relative Rota-Baxter group. It turns out that there is an isomorphism between the two categories under some mild restrictions. We propose an efficient GAP algorithm, which would enable the computation of relative Rota-Baxter operators on finite groups. In the end, we introduce the notion of isoclinism of relative Rota-Baxter groups and prove that an isoclinism of these objects induces an isoclinism of corresponding skew left braces., Comment: Comments and Suggestions are welcome

The notion of Rota-Baxter groups was recently introduced by Guo, Lang and Sheng [{\em Adv. Math.} 387 (2021), 107834, 34 pp.] in the geometric study of Rota-Baxter Lie algebras. They are closely related to skew braces as observed by Bardakov and Gubarev. In this paper, we study extensions of Rota-Baxter groups by constructing suitable cohomology theories. Among others, we find relations with the extensions of skew braces. Given an extension of Rota-Baxter groups, we also construct a short exact sequence connecting various automorphism groups, which generalizes the Wells short exact sequence., Comment: This is a preliminary version of this research. we will soon upload a final version free from typos and errors