arXiv subjects

Kumar Saurav

Publications and source records attributed to Kumar Saurav.

14 recordsLinked to original sources

arXiv2026

Real-Time Voicemail Detection in Telephony Audio Using Temporal Speech Activity Features

Outbound AI calling systems must distinguish voicemail greetings from live human answers in real time to avoid wasted agent interactions and dropped calls. We present a lightweight approach that extracts 15 temporal features from the speech activity pattern of a pre-trained neural voice activity detector (VAD), then classifies with a shallow tree-based ensemble. Across two evaluation sets totaling 764 telephony recordings, the system achieves a combined 96.1% accuracy (734/764), with 99.3% (139/140) on an expert-labeled test set and 95.4% (595/624) on a held-out production set. In production validation over 77,000 calls, it maintained a 0.3% false positive rate and 1.3% false negative rate. End-to-end inference completes in 46 ms on a commodity dual-core CPU with no GPU, supporting 380+ concurrent WebSocket calls. In our search over 3,780 model, feature, and threshold combinations, feature importance was concentrated in three temporal variables. Adding transcription keywords or beep-based features did not improve the best real-time configuration and increased latency substantially. Our results suggest that temporal speech patterns are a strong signal for distinguishing voicemail greetings from live human answers.

cs.SD↗

arXiv2026

Universal Hamiltonian control in a planar trimon circuit

Multimode circuits provide an avenue for flexible control of single and multi-qubit gates. In this work we implement a multimode circuit known as a trimon integrated in a planar geometry. The trimon features three transmon-like modes with strong all-to-all $ZZ$ coupling. We demonstrate high fidelity operations on the trimon, achieving flexible control of its rich state space. This includes qubit rotations conditioned on one or both other qubits, unconditional single-qubit rotations, and both excitation-conserving and double-excitation two-qubit entangling gates. Through multi-tone driving we are able to implement all 16 two-qubit Pauli operators in the two-qubit space. We further demonstrate using the trimon as a qudit with up to 8 states and higher coherence than typical transmon-based implementations. Our results show a compact, highly controllable device that can potentially replace transmons in standard superconducting processor architectures.

quant-ph↗

arXiv2025

Monitoring State Transitions in Markovian Systems with Sampling Cost

We consider a node-monitor pair, where the node's state varies with time. The monitor needs to track the node's state at all times; however, there is a fixed cost for each state query. So the monitor may instead predict the state using time-series forecasting methods, including time-series foundation models (TSFMs), and query only when prediction uncertainty is high. Since query decisions influence prediction accuracy, determining when to query is nontrivial. A natural approach is a greedy policy that predicts when the expected prediction loss is below the query cost and queries otherwise. We analyze this policy in a Markovian setting, where the optimal (OPT) strategy is a state-dependent threshold policy minimizing the time-averaged sum of query cost and prediction losses. We show that, in general, the greedy policy is suboptimal and can have an unbounded competitive ratio, but under common conditions such as identically distributed transition probabilities, it performs close to OPT. For the case of unknown transition probabilities, we further propose a projected stochastic gradient descent (PSGD)-based learning variant of the greedy policy, which achieves a favorable predict-query tradeoff with improved computational efficiency compared to OPT.

cs.LG↗

arXiv2024

Beating the Ramsey limit on sensing with deterministic qubit control

Quantum sensors promise revolutionary advances in medical imaging, energy production, mass detection, geodesy, foundational physics research, and a host of other fields. In many sensors, the signal takes the form of a changing qubit frequency, which is detected with an interference measurement. Unfortunately, environmental noise decoheres the qubit state, reducing signal-to-noise ratio (SNR). Here we introduce a protocol for enhancing the sensitivity of a measurement of a qubit's frequency in the presence of decoherence. We use a continuous drive to stabilize one component of the qubit's Bloch vector, enhancing the effect of a small static frequency shift. We demonstrate our protocol on a superconducting qubit, enhancing SNR per measurement shot by 1.65$\times$ and SNR per qubit evolution time by 1.09$\times$ compared to standard Ramsey interferometry. We explore the protocol theoretically and numerically, finding maximum enhancements of 1.96$\times$ and 1.18$\times$, respectively. We also show that the protocol is robust to parameter miscalibrations. Our protocol provides an unconditional enhancement in signal-to-noise ratio compared to standard Ramsey interferometry. It requires no feedback and no extra control or measurement resources, and can be immediately applied in a wide variety of quantum computing and quantum sensor technologies to enhance their sensitivities.

quant-ph↗

arXiv2024

Quantum Property Preservation

Quantum property preservation (QPP) is the problem of maintaining a target property of a quantum system for as long as possible. This problem arises naturally in the context of open quantum systems subject to decoherence. Here, we develop a general theory to formalize and analyze QPP. We characterize properties encoded as scalar functions of the system state that can be preserved time-locally via continuous control using smoothly varying, time-dependent control Hamiltonians. The theory offers an intuitive geometric interpretation involving the level sets of the target property and the stable and unstable points related to the noise channel. We present solutions for various noise channels and target properties, which are classified as trivially controllable, uncontrollable, or controllable. In the controllable scenario, we demonstrate the existence of control Hamiltonian singularities and breakdown times, beyond which property preservation fails. QPP via Hamiltonian control is complementary to quantum error correction, as it does not require ancilla qubits or rely on measurement and feedback. It is also complementary to dynamical decoupling, since it uses only smooth Hamiltonians without pulsing and works in the regime of Markovian open system dynamics. From the perspective of control theory, this work addresses the challenge of tracking control for open quantum systems.

quant-ph↗

arXiv2024

Benchmarking quantum gates and circuits

Accurate noise characterization in quantum gates and circuits is vital for the development of reliable quantum simulations for chemically relevant systems and fault-tolerant quantum computing. This paper reviews a variety of key benchmarking techniques, including Randomized Benchmarking, Quantum Process Tomography, Gate Set Tomography, Process Fidelity Estimation, Direct Fidelity Estimation, and Cross-Entropy Benchmarking. We evaluate each method's complexities, the resources they require, and their effectiveness in addressing coherent, incoherent, and state preparation and measurement (SPAM) errors. Furthermore, we introduce deterministic benchmarking (DB), a novel protocol that minimizes the number of experimental runs, exhibits resilience to SPAM errors, and effectively characterizes both coherent and incoherent errors. The implementation of DB is experimentally validated using a superconducting transmon qubit, and the results are substantiated with a simple analytical model and master equation simulations. With the addition of DB to the toolkit of available benchmarking methods, this article serves as a practical guide for choosing and applying benchmarking protocols to advance quantum computing technologies.

quant-ph↗

arXiv2023

Minimizing Age of Information under Latency and Throughput Constraints

We consider a scheduling problem pertinent to a base station (BS) that serves heterogeneous users (UEs). We focus on a downlink model, with users that are either interested in large throughput, low latency (the time difference between the packet arrival at the BS, and reception at the UE), or low age-of-information (AoI, the difference between the current time and the arrival time of the latest packet at the BS that has been received at the UE). In each time step, the BS may serve a limited number of UEs. The objective is to find a scheduling algorithm that meets the individual service requirements of all the UEs. In this paper, we formulate the aforementioned objective as an optimization problem, and propose candidate solutions (algorithms) that rationalize the tradeoff between throughput, latency and the AoI. Then, we analyze the performance of the proposed algorithms both analytically and numerically, and discuss techniques to extend their applicability to other general setups.

cs.IT↗

arXiv2022

Minimizing Age of Information under Arbitrary Arrival Model with Arbitrary Packet Size

We consider a single source-destination pair, where information updates arrive at the source at arbitrary time instants. For each update, its size, i.e. the service time required for complete transmission to the destination, is also arbitrary. At any time, age of information (AoI) is equal to the difference between the current time, and the arrival time of the latest update (at the source) that has been completely transmitted (to the destination). AoI quantifies the staleness of the update (information) at the destination. The goal is to find a causal scheduling policy that minimizes the time average of AoI, where the possible decisions at any time are i) whether to preempt the update under transmission upon arrival of a new update, and ii) if no update is under transmission, then choose which update to transmit (among the available updates). In this paper, we propose a causal policy called SRPT$^+$ that at each time, i) preempts the update under transmission if a new update arrives with a smaller size, and ii) if no update is under transmission, then begins to transmit the update for which the ratio of the reduction in AoI upon complete transmission (if not preempted in future) and the remaining size, is maximum. We characterize the performance of SRPT$^+$ using the metric called the competitive ratio, i.e. the ratio of the average AoI of causal policy and the average AoI of an optimal offline policy (that knows the entire input in advance), maximized over all possible inputs. We show that the competitive ratio of SRPT$^+$ is at most $4$. Further, we propose a simpler policy called SRPT$^L$, that i) preempts the update under transmission if a new update arrives with a smaller size, and ii) if no update is under transmission, then begins to transmit the update with the latest arrival time. We show that the competitive ratio of SRPT$^L$ is at most $29$.

cs.IT↗

arXiv2022

Scheduling to Minimize Age of Information with Multiple Sources

We consider a G/G/1 queueing system with a single server, where updates arrive from different sources stochastically with possibly different update inter-generation time distributions. The server can transmit/serve at most one update at any time, with potentially different transmission/service times for updates belonging to distinct sources. The age of information (AoI) of any source is a function of the time difference between the departure time of successive updates of that source. Each fully/partially transmitted update incurs a fixed (energy) cost, and the goal of the scheduler is to minimize the linear combination of the sum of the age of information across all sources and the total energy cost. We propose a simple non-preemptive randomized scheduling algorithm that randomly marks arriving updates from a source to be eligible for transmission with a fixed probability and discards them otherwise. Every time the server becomes free, it chooses a source for transmission randomly with another fixed probability and begins to transmit the most recently marked update of the chosen source. Both the respective probabilities are chosen by solving a convex program. The competitive ratio of the proposed algorithm (against a non-preemptive offline optimal algorithm) is shown to be 3 plus the maximum of the ratio of the variance and the mean of the inter-arrival time distribution of sources. For several common distributions such as exponential, uniform and Rayleigh, the competitive ratio is at most 4. For preemptive policies, a G/M/1 system is considered and a non-preemptive policy is shown to have competitive ratio (against a preemptive offline optimal algorithm) at most 5 plus the maximum of the ratio of the variance and the mean of the inter-arrival time distribution of sources.

cs.IT↗

arXiv2022

3-Competitive Policy for Minimizing Age of Information in Multi-Source M/G/1 Queuing Model

We consider a multi-source network with a common monitor, where fresh updates are generated at each source, following a Poisson process. At any time, at most one source can transmit its update to the monitor, and transmission time for updates of each source follows some general distribution. The goal is to find a causal scheduling policy such that at any time, the latest update available at each source is fresh. In this paper, we quantify freshness using the age of information (AoI) metric, and propose a randomized policy, which we show is 3-competitive with respect to Pareto-optimal policies (that minimize the expected average AoI of each source). We also show that for a particular choice of the randomization parameter, the proposed randomized policy is 3-competitive with respect to an optimal policy that minimizes the weighted sum of the expected average AoI of all sources.

cs.IT↗

arXiv2021

"A Passage to India": Pre-trained Word Embeddings for Indian Languages

Dense word vectors or 'word embeddings' which encode semantic properties of words, have now become integral to NLP tasks like Machine Translation (MT), Question Answering (QA), Word Sense Disambiguation (WSD), and Information Retrieval (IR). In this paper, we use various existing approaches to create multiple word embeddings for 14 Indian languages. We place these embeddings for all these languages, viz., Assamese, Bengali, Gujarati, Hindi, Kannada, Konkani, Malayalam, Marathi, Nepali, Odiya, Punjabi, Sanskrit, Tamil, and Telugu in a single repository. Relatively newer approaches that emphasize catering to context (BERT, ELMo, etc.) have shown significant improvements, but require a large amount of resources to generate usable models. We release pre-trained embeddings generated using both contextual and non-contextual approaches. We also use MUSE and XLM to train cross-lingual embeddings for all pairs of the aforementioned languages. To show the efficacy of our embeddings, we evaluate our embedding models on XPOS, UPOS and NER tasks for all these languages. We release a total of 436 models using 8 different approaches. We hope they are useful for the resource-constrained Indian language NLP. The title of this paper refers to the famous novel 'A Passage to India' by E.M. Forster, published initially in 1924.

cs.CL↗

arXiv2021

Minimizing the Sum of Age of Information and Transmission Cost under Stochastic Arrival Model

We consider a node-monitor pair, where updates are generated stochastically (according to a known distribution) at the node that it wishes to send to the monitor. The node is assumed to incur a fixed cost for each transmission, and the objective of the node is to find the update instants so as to minimize a linear combination of AoI of information and average transmission cost. First, we consider the Poisson arrivals case, where updates have an exponential inter-arrival time for which we derive an explicit optimal online policy. Next, for arbitrary distributions of inter-arrival time of updates, we propose a simple randomized algorithm that transmits any newly arrived update with a fixed probability (that depends on the distribution) or never transmits that update. The competitive ratio of the proposed algorithm is shown to be a function of the variance and the mean of the inter-arrival time distribution. For some of the commonly considered distributions such as exponential, uniform, and Rayleigh, the competitive ratio bound is shown to be 2.

cs.IT↗

arXiv2021

Online Energy Minimization Under A Peak Age of Information Constraint

We consider a node where packets of fixed size (in bits) are generated at arbitrary intervals. The node is required to maintain the peak age of information (AoI) at the monitor below a threshold by transmitting potentially a subset of the generated packets. At any time, depending on the packet availability and the current AoI, the node can choose which packet to transmit, and at what transmission speed (in bits per second). Power consumption is a monotonically increasing convex function of the transmission speed. In this paper, for any given time horizon, the objective is to find a causal policy that minimizes the total energy consumption while satisfying the peak AoI constraint. We consider competitive ratio as the performance metric, that is defined as the ratio of the expected cost of a causal policy, and the expected cost of an optimal offline policy that knows the input (packet generation times) in advance. We first derive a lower bound on the competitive ratio of all causal policies, in terms of the system parameters (such as power function, packet size and peak AoI threshold), and then propose a particular policy for which we show that its competitive ratio has similar order of dependence on the system parameters as the derived lower bound.

cs.IT↗

arXiv2020

Game of Ages

We consider a distributed IoT network, where each node wants to minimize its age of information and there is a cost to make any transmission. A collision model is considered, where any transmission is successful from a node to a common monitor if no other node transmits in the same slot. There is no explicit communication/coordination between any two nodes. The selfish objective of each node is to minimize a function of its individual age of information and its transmission cost. Under this distributed competition model, the objective of this paper is to find a distributed transmission strategy for each node that converges to an equilibrium. The proposed transmission strategy only depends on the past observations seen by each node and does not require explicit information of the number of other nodes, or their strategies. A simple update strategy is shown to converge to an equilibrium, that is in fact a Nash equilibrium for a suitable utility function, that captures all the right tradeoffs for each node. In addition, the price of anarchy for the utility function is shown to approach unity as the number of nodes grows large.

cs.IT↗