arXiv subjects

Yash Deshpande

Publications and source records attributed to Yash Deshpande.

At least 19 recordsLinked to original sources

arXiv2026

Strict-Priority Packet Delay in Switches with Transmit-Ring Buffering

Strict Priority (SP) scheduling is widely used at switch egress to provide low-latency service to high-priority (HP) traffic. Existing deterministic and stochastic latency models typically account for scheduler behavior and packet transmission, but omit a common switch implementation detail: the transmit ring (TXR) between the scheduler and the physical port. Because the switch must prepare the next packet before the current transmission completes, packets already placed in the TXR can further delay HP packets. This changes both the worst-case delay and the per-hop delay distribution of HP packets. This paper identifies this modeling gap, extends standard SP latency models to include the TXR, and validates the revised model through measurements on multiple switches. It also provides a measurement method for estimating the TXR size, a parameter that is often not reported in switch datasheets. The resulting model provides a closer representation of switch behavior for systems that use SP scheduling and require either delay bounds or delay distributions.

cs.NI↗

arXiv2026

Pandora's Regret: A Proper Scoring Rule for Evaluating Sequential Search

In sequential search, alternatives are tested until the true class is found. Standard proper scoring rules like log loss are local, ignoring the ranking of competitors and misaligning model evaluation with search utility. We show that sequential search induces a pairwise structure that overcomes this. By analyzing the expected cost of optimal search under varying testing costs, we derive Pandora's Regret: a closed-form, pairwise-additive, and strictly proper scoring rule. Pandora's Regret both elicits true probabilities and penalizes rank-reversing miscalibrations where distractors outrank the true class. Our construction yields a one-parameter Beta family that balances penalties for rank-swapping versus probability magnitude, while retaining a grounded interpretation as expected search cost. We prove that log loss, accuracy, and macro-F1 rely on implicit decision models misaligned with sequential search. Across 597 MedMNIST models, Pandora-based metrics better predict clinical diagnostic costs than standard alternatives, extending decision-theoretic scoring rule construction to the multiclass setting.

cs.LG↗

arXiv2026

Improving Network Clock Synchronization by Marking Congestion

Achieving consistent time across devices in distributed systems often involves exchanging timestamped messages over a network. Precise time synchronization is crucial for applications such as cellular networks, industrial automation, and transactional databases. However, delay variation in synchronization packets-often caused by congestion from competing traffic-degrades synchronization accuracy. Detecting whether a packet experienced congestion can help improve synchronization through filtering and statistical methods. We propose an in-network congestion indication and filtering mechanism for synchronization messages used in protocols such as the Network Time Protocol (NTP) and Precision Time Protocol (PTP). Network devices mark packets that experienced queuing, allowing clocks to correct errors caused by varying delays. Our approach requires only simple changes at switches or routers, avoiding deep packet inspection or protocol modifications. The method is backward compatible, using standard but currently unused fields in IP, PTP, or NTP headers. We implement our method on a Tofino P4 target and demonstrate an improvement of over 80% in synchronization performance over a single hop. Moreover, we show that the performance of traditional statistical filters, such as min-RTT and median-delay, is improved by 90% over the one-hop hardware setup. We further demonstrate the effectiveness of our proposed method across multiple hops, both analytically and through simulation. Congestion marking improves the root-mean-squared clock offset estimation error by 30% to 80%, depending on network conditions and filtering techniques.

cs.NI↗

arXiv2026

Vulnerability Analysis of eBPF-enabled Containerized Deployments of 5G Core Networks

The extended Berkeley Packet Filter (eBPF) is useful for faster packet processing and network monitoring in softwarized deployments. Similarly, softwarized deployments of 5G core network services adopted eBPF to meet the stringent latency and bandwidth requirements of underlying applications. While the existing studies focused on network performance, security concerns over eBPF-enabled platforms are overlooked. In this paper, we study the vulnerability analysis of 5G core network deployments that use eBPF for packet processing and traffic monitoring. In particular, we consider the following aspects: a) tracing, b) denial-of-service (DoS), c) stealing information, and d) bash injection. We present the detailed attack scenarios with step-by-step implementation of containerized and eBPF-enabled 5G network functions using Open5GS. The experiment results show that the aforementioned vulnerabilities are present in eBPF-enabled 5G deployments and can be exploited by attackers. Finally, we present some mitigation techniques useful for addressing the vulnerabilities. The source code and implementation details are made available at https://github.com/chimms1/5G-eBPF-exploits.

cs.NI↗

arXiv2025

Joint Resource Allocation to Transparently Integrate 5G TDD Uplink with Time-Aware TSN

To enable mobility in industrial communication systems, the seamless integration of 5G with Time-Sensitive Networking (TSN) is a promising approach. Deterministic communication across heterogeneous 5G-TSN systems requires joint scheduling between both domains. A key prerequisite for time-aware end-to-end scheduling is determining the forwarding delay for each TSN Traffic Class at every bridge, referred to as Bridge Delay (BD). Hence, to integrate 5G as a transparent TSN bridge, the 5G BD must be determined and guaranteed. Unlike wired bridges, the 5G BD relies on wireless resource management characteristics, such as the Time Division Duplex pattern and radio resource allocation procedure. In particular, traditional Uplink (UL) schedulers are optimized for throughput but often fail to meet the deadline requirements. To address this challenge, we propose a heterogeneous radio resource scheduler that integrates static and dynamic scheduling. The algorithm pre-allocates resources for time-sensitive periodic streams based on the reported BDs, ensuring alignment with the TSN mechanisms Time-Aware Shaper and Per-Stream Filtering and Policing. Meanwhile, remaining resources are dynamically allocated to non-deterministic flows using established strategies such as Proportional Fair, Max C/I, or a Quality of Service-aware priority-based scheduler. The scheduler's performance is evaluated through OMNeT++ simulations. The results demonstrate support for diverse TSN flows while ensuring deadline-aware scheduling of time-sensitive UL traffic in mobility scenarios. Periodic time-sensitive flows are end-to-end scheduled across domains, improving the resource efficiency by 28% compared to the Configured Grant baseline. While reliability is preserved, non-deterministic rate-sensitive flows benefit from the improved resource utilization, resulting in higher throughput

cs.NI↗

arXiv2025

Predicting ulcer in H&E images of inflammatory bowel disease using domain-knowledge-driven graph neural network

Inflammatory bowel disease (IBD) involves chronic inflammation of the digestive tract, with treatment options often burdened by adverse effects. Identifying biomarkers for personalized treatment is crucial. While immune cells play a key role in IBD, accurately identifying ulcer regions in whole slide images (WSIs) is essential for characterizing these cells and exploring potential therapeutics. Multiple instance learning (MIL) approaches have advanced WSI analysis but they lack spatial context awareness. In this work, we propose a weakly-supervised model called DomainGCN that employs a graph convolution neural network (GCN) and incorporates domain-specific knowledge of ulcer features, specifically, the presence of epithelium, lymphocytes, and debris for WSI-level ulcer prediction in IBD. We demonstrate that DomainGCN outperforms various state-of-the-art (SOTA) MIL methods and show the added value of domain knowledge.

eess.IV↗

arXiv2025

TwinRAN: Twinning the 5G RAN in Azure Cloud

The proliferation of 5G technology necessitates advanced network management strategies to ensure optimal performance and reliability. Digital Twin (DT)s have emerged as a promising paradigm for modeling and simulating complex systems like the 5G Radio Access Network (RAN). In this paper, we present TwinRAN, a DT of the 5G RAN built leveraging the Azure DT platform. TwinRAN is built on top of the Open RAN (O-RAN) architecture and is agnostic to the vendor of the underlying equipment. We demonstrate three applications using TwinRAN and evaluate the required resources and their performance for a network with 800 users and eight gNBs. We first evaluate the performance and limitations of the Azure DT platform, measuring the latency under different conditions. The results from this evaluation allow us to optimize TwinRAN for the DT platform it uses. Then, we present the system's architectural design, emphasizing its components and interactions. We propose that two types of twin graphs be simultaneously maintained on the cloud: one for intercell operations, keeping a broad overview of all the cells in the network, and another where each cell is spawned in a separate Azure DT instance for more granular operation and monitoring of intracell tasks. We evaluate the performance and operating costs of TwinRAN for each of the three applications. The TwinRAN DT in the cloud can keep track of its physical twin within a few hundred milliseconds, extending its utility to many 5G network management tasks, some of which are shown in this paper. The novel framework for building and maintaining a DT of the 5G RAN presented in this paper offers network operators enhanced capabilities, empowering efficient deployments and management.

cs.NI↗

arXiv2024

Integrating Deterministic Networking with 5G

The rising prevalence of real-time applications that require deterministic communication over mobile networks necessitates the joint operation of both mobile and fixed network components. This joint operation requires designing components that interact between the two technologies to provide users with latency and packet loss guarantees. In this work, we demonstrate a fully integrated 5G-DetNet that can guarantee the end-to-end demands of different flows. Moreover, we show how such a network can be implemented using low-cost hardware and open-source software, making it accessible to many 5G testbeds. The features demonstrated in this work are a network manager that does the routing and scheduling, an application function in the 5G core that interfaces with the network manager, and a network-side translator for user-plane management and de-jittering of the real-time streams.

cs.NI↗

arXiv2024

LCDN: Providing Network Determinism with Low-Cost Switches

The demands on networks are increasing at a fast pace. In particular, real-time applications have very strict network requirements. However, building a network that hosts real-time applications is a cost-intensive endeavor, especially for experimental systems such as testbeds. Systems that provide guaranteed real-time networking capabilities usually work with expensive software-defined switches. In contrast, real-time networking systems based on low-cost hardware face the limitation of lower link speeds. This paper fills this gap and presents Low-Cost Deterministic Networking (LCDN), a system designed to work with inexpensive, common off-the-shelf switches and devices. LCDN works at Gigabit speed and enables powerful testbeds to host real-time applications with strict delay guarantees. This paper also provides an evaluation of the determinism of the switch and a Raspberry Pi used as an end device to demonstrate the applicability of LCDN on inexpensive low-power reduced capacity apparatus.

cs.NI↗

arXiv2024

OpenAirLink: Reproducible Wireless Channel Emulation using Software Defined Radios

This paper presents OpenAirLink(OAL), an open-source channel emulator for reproducible testing of wireless scenarios. OAL is implemented on off-the-shelf software-defined radios (SDR) and presents a smaller-scale alternative to expensive commercially available channel emulators. Path loss and propagation delay are the fundamental aspects of emulating a wireless channel. OAL provides a simple method to change these aspects in real-time. The emulator is implemented using a finite impulse response (FIR) filter. The FIR filter is written in Verilog and flashed on the SDRs Field Programmable Gate Array (FPGA). Most processing transpires on the FPGA, so OAL does not require high-performance computing hardware and SDRs. We validate the performance of OAL and demonstrate the utility of such a channel emulation tool using two examples. We believe that open-source channel emulators such as OAL can make reproducible wireless experiments accessible to many researchers in the scientific community.

cs.NI↗

arXiv2023

An Advanced Tree Algorithm with Interference Cancellation in Uplink and Downlink

In this paper, we propose Advanced Tree-algorithm with Interference Cancellation (ATIC), a variant of binary tree-algorithm with successive interference cancellation (SICTA) introduced by Yu and Giannakis. ATIC assumes that Interference Cancellation (IC) can be performed both by the access point (AP), as in SICTA, but also by the users. Specifically, after every collision slot, the AP broadcasts the observed collision as feedback. Users who participated in the collision then attempt to perform IC by subtracting their transmissions from the collision signal. This way, the users can resolve collisions of degree 2 and, using a simple distributed arbitration algorithm based on user IDs, ensure that the next slot will contain just a single transmission. We show that ATIC reaches the asymptotic throughput of 0.924 as the number of initially collided users tends to infinity and reduces the number of collisions and packet delay. We also compare ATIC with other tree algorithms and indicate the extra feedback resources it requires.

cs.IT↗

arXiv2023

Evaluation of NR-Sidelink for Cooperative Industrial AGVs

Industry 4.0 has brought to attention the need for a connected, flexible, and autonomous production environment. The New Radio (NR)-sidelink, which was introduced by the third-generation partnership project (3GPP) in Release 16, can be particularly helpful for factories that need to facilitate cooperative and close-range communication. Automated Guided Vehicles (AGVs) are important for material handling and carriage within these environments, and using NR-sidelink communication can further enhance their performance. An efficient resource allocation mechanism is required to ensure reliable communication and avoid interference between AGVs and other wireless systems in the factory using NR-sidelink. This work evaluates the 3GPP standardized resource allocation algorithm for NR-sidelink for a use case of cooperative carrying AGVs. We suggest further improvements that are tailored to the quality of service (QoS) requirements of an indoor factory communication scenario with cooperative AGVs.The use of NR-sidelink communication has the potential to help meet the QoS requirements for different Industry 4.0 use cases. This work can be a foundation for further improvements in NR-sidelink in 3GPP Release 18 and beyond.

cs.NI↗

arXiv2023

Agreement and Statistical Efficiency in Bayesian Perception Models

Bayesian models of group learning are studied in Economics since the 1970s. and more recently in computational linguistics. The models from Economics postulate that agents maximize utility in their communication and actions. The Economics models do not explain the ``probability matching" phenomena that are observed in many experimental studies. To address these observations, Bayesian models that do not formally fit into the economic utility maximization framework were introduced. In these models individuals sample from their posteriors in communication. In this work we study the asymptotic behavior of such models on connected networks with repeated communication. Perhaps surprisingly, despite the fact that individual agents are not utility maximizers in the classical sense, we establish that the individuals ultimately agree and furthermore show that the limiting posterior is Bayes optimal. We explore the interpretation of our results in terms of Large Language Models (LLMs). In the positive direction our results can be interpreted as stating that interaction between different LLMs can lead to optimal learning. However, we provide an example showing how misspecification may lead LLM agents to be overconfident in their estimates.

math.ST↗

arXiv2023

Near-optimal inference in adaptive linear regression

When data is collected in an adaptive manner, even simple methods like ordinary least squares can exhibit non-normal asymptotic behavior. As an undesirable consequence, hypothesis tests and confidence intervals based on asymptotic normality can lead to erroneous results. We propose a family of online debiasing estimators to correct these distributional anomalies in least squares estimation. Our proposed methods take advantage of the covariance structure present in the dataset and provide sharper estimates in directions for which more information has accrued. We establish an asymptotic normality property for our proposed online debiasing estimators under mild conditions on the data collection process and provide asymptotically exact confidence intervals. We additionally prove a minimax lower bound for the adaptive linear regression problem, thereby providing a baseline by which to compare estimators. There are various conditions under which our proposed estimators achieve the minimax lower bound. We demonstrate the usefulness of our theory via applications to multi-armed bandit, autoregressive time series estimation, and active learning with exploration.

math.ST↗

arXiv2023

Analysis of d-ary Tree Algorithms with Successive Interference Cancellation

In this article, we calculate the mean throughput, number of collisions, successes, and idle slots for random tree algorithms with successive interference cancellation. Except for the case of the throughput for the binary tree, all the results are new. We furthermore disprove the claim that only the binary tree maximises throughput. Our method works with many observables and can be used as a blueprint for further analysis.

math.PR↗

arXiv2022

Tree-Algorithms with Multi-Packet Reception and Successive Interference Cancellation

In this paper, we perform a thorough analysis of tree-algorithms with multi-packet reception (MPR) and successive interference cancellation (SIC), showing a number of novel results. We first derive the basic performance parameters, which are the expected length of the collision resolution interval and the normalized throughput, conditioned on the number of contending users. We then study their asymptotic behaviour, identifying an oscillatory component that amplifies with the increase in MPR. In the next step, we derive the throughput for the gated and windowed access, assuming Poisson arrivals. We show that for windowed access, the bound on maximum stable normalized throughput increases with the increase in MPR. his implies that investing in advanced physical capabilities, i.e., MPR and SIC pays off from the perspective of the medium access control algorithm.

cs.IT↗

arXiv2022

On d-ary tree algorithms with successive interference cancellation

In this paper, we outline the approach for the derivation of the length of the collision resolution interval for d-ary tree algorithms (TA) with gated access and successive interference cancellation (SIC), conditioned on the number of the contending users. This is the basic performance parameter for TA with gated access. We identify the deficiencies of the analysis performed in the seminal paper on TA with SIC by Yu and Giannakis, showing that their analysis is correct only for binary splitting, i.e. for d=2. We also provide some insightful results on the stable throughput that can be achieved for different values of d.

cs.IT↗

arXiv2022

Improving AoI via Learning-based Distributed MAC in Wireless Networks

In this work, we consider a remote monitoring scenario in which multiple sensors share a wireless channel to deliver their status updates to a process monitor via an access point (AP). Moreover, we consider that the sensors randomly arrive and depart from the network as they become active and inactive. The goal of the sensors is to devise a medium access strategy to collectively minimize the long-term mean network \ac{AoI} of their respective processes at the remote monitor. For this purpose, we propose specific modifications to ALOHA-QT algorithm, a distributed medium access algorithm that employs a policy tree (PT) and reinforcement learning (RL) to achieve high throughput. We provide the upper bound on the mean network Age of Information (AoI) for the proposed algorithm along with pointers for selecting its key parameter. The results reveal that the proposed algorithm reduces mean network \ac{AoI} by more than 50 percent for state of the art stationary randomized policies while successfully adjusting to a changing number of active users in the network. The algorithm needs less memory and computation than ALOHA-QT while performing better in terms of AoI.

cs.NI↗