arXiv subjects

Arpit Gupta

Publications and source records attributed to Arpit Gupta.

At least 19 recordsLinked to original sources

arXiv2026

Pramana: A Composable, Domain-Specific Backend for Empirical Networking Research

Networking research advances by turning hypotheses into empirical evidence, so accelerating it means reducing the lag between ideation (synthesizing a hypothesis) and generating the data that tests it. Consider a concrete case: does a bulk BBR download fairly share its bottleneck with competing real-time Google Meet traffic? Validating this requires configuring a realistic bottleneck link, concurrently generating BBR's bulk transfer and Meet's real-time traffic, and collecting relevant service-quality metrics. Today this overhead is high, often forcing researchers to start from scratch for every new idea. This ideation-to-data-generation gap will only worsen in the agentic AI era, where AI-assisted ideation accelerates exponentially, yet its outputs cannot be validated without a data-generation backend. This paper explores how to bridge this gap. We envision a composable, domain-specific backend, Pramana, shaped as a thin waist, with diverse research intents at the top and disparate execution substrates at the bottom. Pramana realizes this waist through a single contract, the intent specification, which disaggregates an experiment into three independent axes: the intent (what data to generate), the substrate (where to generate it), and the mechanism (how to produce it), so one specification runs on any substrate. We demonstrate Pramana's utility by building a first-of-its-kind corpus of 255 data-generation intents mined from 66 published papers, and show the intent specification satisfies all of them, where no existing tool satisfies more than 13%. Our current proof-of-concept implementation already satisfies 34% of these intents, more than twice the best existing tool, and we lay out a roadmap for closing this abstraction-implementation gap through a broader community effort to build the envisioned data-generation backend and accelerate empirical networking research.

cs.NI↗

arXiv2026

Robust and Extensible Measurement of Broadband Plans with BQT+

Independent, street address-level broadband data is essential for evaluating Internet infrastructure investments, such as the $42B Broadband Equity, Access, and Deployment (BEAD) program. Evaluating these investments requires longitudinal visibility into broadband availability, quality, and affordability, including data on pre-disbursement baselines and changes in providers' advertised plans. While such data can be obtained through Internet Service Provider (ISP) web interfaces, these workloads impose three fundamental system requirements: robustness to frequent interface evolution, extensibility across hundreds of providers, and low technical overhead for non-expert users. Existing systems fail to meet these three essential requirements. We present BQT+, a broadband plan measurement framework that replaces monolithic workflows with declarative state/action specifications. BQT+ models querying intent as an interaction state space, formalized as an abstract nondeterministic finite automaton (NFA), and selects execution paths at runtime to accommodate alternative interaction flows and localized interface changes. We show that BQT+ sustains longitudinal monitoring of 64 ISPs, supporting querying for over 100 ISPs. We apply it to two policy studies: constructing a BEAD pre-disbursement baseline and benchmarking broadband affordability across over 124,000 addresses in four states.

cs.NI↗

arXiv2025

ScalePRM: Training Process Reward Models by Scaling Verification Compute Without Ground Truth

Training process reward models (PRMs) requires step-level correctness labels, obtained either through expensive human annotation or by relying on ground-truth answers, limiting the ability to scale process-level supervision. We propose ScalePRM, which scales verification compute as an alternative: given a problem and a candidate solution, we generate multiple independent verifications of each reasoning step and aggregate their judgments to produce synthetic step-level labels without ground truth. We explore two representative inference-time scaling strategies, parallel scaling through self-consistency and sequential scaling through meta-critique, and train generative PRMs on the resulting synthetic data. On ProcessBench, a benchmark for identifying erroneous steps in mathematical reasoning, PRMs trained on step-level self-consistency data achieve 67.5 F1, surpassing reference-guided training with ground-truth access (66.4 F1) and GPT-4o as a critic (61.9 F1). When deployed as reward signals in RL training with Qwen2.5-Math-7B, our best PRM achieves 47.4% average accuracy across six mathematical reasoning benchmarks, outperforming ground-truth-based RLVR (43.9%). We also identify and address reward exploitation patterns unique to generative PRM-based RL. Our results demonstrate that scaling verification compute is a viable alternative to ground-truth supervision for training process reward models.

cs.LG↗

arXiv2025

SABER: Small Actions, Big Errors -- Safeguarding Mutating Steps in LLM Agents

Despite rapid progress in LLM agents, performance on long-horizon, tool-using tasks remains fragile. To better understand this fragility, we ask a simple question: \emph{do all actions contribute equally to failure?} Analyzing execution traces on $\tau$-Bench (Airline/Retail) and SWE-Bench Verified, we decompose trajectories into \emph{mutating} (environment-changing) vs.\ non-mutating steps and formalize \emph{decisive deviations}, earliest action, level divergences that flip success to failure. A logistic regression reveals that each additional deviation in a mutating action reduces the odds of success by upto $92\%$ on Airline and upto $96\%$ on Retail for SoTA models. In contrast, deviations in non-mutating actions have little to no effect. Errors also grow with context length as agents drift from role and act on stale constraints. Motivated by these observations, we introduce \cm{}, a model-agnostic, gradient-free, test-time safeguard that (i) adds mutation-gated verification, (ii) injects \emph{Targeted Reflection} before mutating steps, and (iii) performs block-based context cleaning. \cm{} delivers consistent gains, e.g., Qwen3-Thinking: +28\% \emph{relative} on Airline, +11\% on Retail, and +7\% on SWE-Bench Verified; Claude: +9\%/+7\%. We further identify ceiling effects in $\tau$-Bench, where annotation errors and underspecified tasks artificially cap model performance. To address this, we release $\tau$-Bench Verified, which restores benchmark headroom through targeted revisions. Our results argue for action-level analysis, targeted safeguards, and reliable evaluations as prerequisites for robust multi-turn agents.

cs.LG↗

arXiv2025

Enabling Data-Driven Policymaking Using Broadband-Plan Querying Tool (BQT+)

Poor broadband access undermines civic and economic life, a challenge exacerbated by the fact that millions of Americans still lack reliable high-speed connectivity. Federal broadband funding initiatives aim to address these gaps, but their success depends on accurate availability and affordability data. Existing data, often based on self-reported ISP information, can overstate coverage and speeds, risking misallocation of funds and leaving unserved communities behind. We present BQT+, an AI-agent data collection platform that queries ISP web interfaces by inputting residential street addresses and extracting data on service availability, quality, and pricing. BQT+ has been used in policy evaluation studies, including an independent assessment of broadband availability, speed tiers, and affordability in areas targeted by the $42.45 billion BEAD program.

cs.NI↗

arXiv2025

NetBurst: Event-Centric Forecasting of Bursty, Intermittent Time Series

Network operators monitor their infrastructure by collecting telemetry data such as packet counts, byte rates, or flow volumes, yet answering the questions that effective operations demand -- forecasting future load, diagnosing and characterizing anomalies, and searching for and retrieving historical precedents -- requires more than raw measurements. Bridging this gap calls for learned representations: compact per-entity summaries that capture temporal dynamics from each entity's univariate time series. Time-series foundation models are the natural starting point, but they are designed for dense, periodic benchmark datasets -- the \emph{mild} statistical regime. However, network telemetry data inhabits the \emph{wild} regime: operationally relevant events are rare, separated by variable-length stretches of low or no activity (``ebbs''), with intermittent bursts of heavy-tailed extremes (``tides''). We present NetBurst, an event-centric pipeline that collapses ebbs, separates each time series into a stream of burst timings and a stream of burst magnitudes, and learns a single representation serving all three operational tasks. Compared to the strongest competitors among eight baselines -- including Amazon's Chronos-2 and Datadog's Toto -- and across nine production telemetry configurations, NetBurst reduces median forecasting error by $1.3$--$116\times$ on wild-regime data with a $1.0$--$7.5\times$ better match to the true burst distribution, and matches baselines on mild-regime benchmarks. For characterizing anomalies, NetBurst produces balanced, well-spread clusters that are $16\times$ more describable in operator-familiar terms under a novel interpretability score, and cluster-filtered search delivers $7.5\times$ faster end-to-end retrieval.

cs.NI↗

arXiv2025

TURBOTEST: Learning When Less is Enough through Early Termination of Internet Speed Tests

Internet speed tests are indispensable for users, ISPs, and policymakers, but their static flooding-based design imposes growing costs: a single high-speed test can transfer hundreds of MB, and collectively, platforms like Ookla, M-Lab, and Fast.com generate petabytes of traffic each month. Reducing this burden requires deciding when a test can be stopped early without sacrificing accuracy. We frame this as an optimal stopping problem and show that existing heuristics-static thresholds, BBR pipe-full signals, or throughput stability rules from Fast.com and FastBTS-capture only a narrow slice of the achievable accuracy-savings trade-off. This paper introduces TurboTest, a systematic framework for speed test termination that sits atop existing platforms. The key idea is to decouple throughput prediction (Stage 1) from test termination (Stage 2): Stage 1 trains a regressor to estimate final throughput from partial measurements, while Stage 2 trains a classifier to decide when sufficient evidence has accumulated to stop. Leveraging richer transport-level features (RTT, retransmissions, congestion window) alongside throughput, TurboTest exposes a single tunable parameter epsilon for accuracy tolerance and includes a fallback mechanism for high-variability cases. Evaluation on 1 million M-Lab NDT speed tests (2024-2025) shows that TurboTest achieves 1.8-4.4x higher data savings than an approach based on BBR signals while reducing median error. These results demonstrate that adaptive ML-based termination can deliver accurate, efficient, and deployable speed tests at scale.

cs.NI↗

arXiv2025

Demystifying Network Foundation Models

This work presents a systematic investigation into the latent knowledge encoded within Network Foundation Models (NFMs) that focuses on hidden representations analysis rather than pure downstream task performance. Different from existing efforts, we analyze the models through a three-part evaluation: Embedding Geometry Analysis to assess representation space utilization, Metric Alignment Assessment to measure correspondence with domain-expert features, and Causal Sensitivity Testing to evaluate robustness to protocol perturbations. Using five diverse network datasets spanning controlled and real-world environments, we evaluate four state-of-the-art NFMs, revealing that they all exhibit significant anisotropy, inconsistent feature sensitivity patterns, an inability to separate the high-level context, payload dependency, and other properties. Our work identifies numerous limitations across all models and demonstrates that addressing them can significantly improve model performance (by up to +0.35 $F_1$ score without architectural changes).

cs.LG↗

arXiv2025

SpliDT: Partitioned Decision Trees for Scalable Stateful Inference at Line Rate

Machine learning (ML) is increasingly being deployed in programmable data planes (switches and SmartNICs) to enable real-time traffic analysis, security monitoring, and in-network decision-making. Decision trees (DTs) are particularly well-suited for these tasks due to their interpretability and compatibility with data-plane architectures, i.e., match-action tables (MATs). However, existing in-network DT implementations are constrained by the need to compute all input features upfront, forcing models to rely on a small, fixed set of features per flow. This significantly limits model accuracy and scalability under stringent hardware resource constraints. We present SPLIDT, a system that rethinks DT deployment in the data plane by enabling partitioned inference over sliding windows of packets. SPLIDT introduces two key innovations: (1) it assigns distinct, variable feature sets to individual sub-trees of a DT, grouped into partitions, and (2) it leverages an in-band control channel (via recirculation) to reuse data-plane resources (both stateful registers and match keys) across partitions at line rate. These insights allow SPLIDT to scale the number of stateful features a model can use without exceeding hardware limits. To support this architecture, SPLIDT incorporates a custom training and design-space exploration (DSE) framework that jointly optimizes feature allocation, tree partitioning, and DT model depth. Evaluation across multiple real-world datasets shows that SPLIDT achieves higher accuracy while supporting up to 5x more stateful features than prior approaches (e.g., NetBeacon and Leo). It maintains the same low time-to-detection (TTD) as these systems, while scaling to millions of flows with minimal recirculation overhead (<0.05%).

cs.NI↗

arXiv2025

NetGent: Agent-Based Automation of Network Application Workflows

We present NetGent, an AI-agent framework for automating complex application workflows to generate realistic network traffic datasets. Developing generalizable ML models for networking requires data collection from network environments with traffic that results from a diverse set of real-world web applications. However, using existing browser automation tools that are diverse, repeatable, realistic, and efficient remains fragile and costly. NetGent addresses this challenge by allowing users to specify workflows as natural-language rules that define state-dependent actions. These abstract specifications are compiled into nondeterministic finite automata (NFAs), which a state synthesis component translates into reusable, executable code. This design enables deterministic replay, reduces redundant LLM calls through state caching, and adapts quickly when application interfaces change. In experiments, NetGent automated more than 50+ workflows spanning video-on-demand streaming, live video streaming, video conferencing, social media, and web scraping, producing realistic traffic traces while remaining robust to UI variability. By combining the flexibility of language-based agents with the reliability of compiled execution, NetGent provides a scalable foundation for generating the diverse, repeatable datasets needed to advance ML in networking.

cs.AI↗

arXiv2025

FLAMES: Improving LLM Math Reasoning via a Fine-Grained Analysis of the Data Synthesis Pipeline

Recent works improving LLM math reasoning with synthetic data have used unique setups, making comparison of data synthesis strategies impractical. This leaves many unanswered questions about the roles of different factors in the synthetic data pipeline, such as the impact of filtering low-quality problems. To address this gap, we introduce FLAMES, a Framework for LLM Assessment of Math rEasoning Data Synthesis, and perform a systematic study of 10 existing data synthesis strategies and multiple other factors impacting the performance of synthetic math reasoning data. Our FLAMES experiments provide several valuable insights about the optimal balance of difficulty and diversity of synthetic data. First, data agents designed to increase problem complexity lead to best improvements on most math metrics. Second, with a fixed data generation budget, keeping higher problem coverage is more important than keeping only problems with reliable solutions. Third, GSM8K- and MATH-based synthetic data can lead to improvements on competition-level benchmarks, showcasing easy-to-hard generalization. Leveraging insights from our FLAMES experiments, we design two novel data synthesis strategies for improving out-of-domain generalization and robustness. Further, we develop the FLAMES dataset, an effective blend of our novel and existing data synthesis strategies, outperforming public datasets on OlympiadBench (+15.7), CollegeMath (+4.5), GSMPlus (+6.5), and MATH (+3.1). Fine-tuning Qwen2.5-Math-7B on the FLAMES dataset achieves 81.4% on MATH, surpassing larger Llama3 405B, GPT-4o and Claude 3.5 Sonnet.

cs.LG↗

arXiv2025

NetForge: A Programmable Substrate for Bottleneck-Centric Network Data Generation

The behavior of Internet applications is shaped by congestion dynamics at bottleneck links, yet data capturing application behavior across diverse bottleneck regimes remains scarce. Bridging this gap requires a data-generation substrate that simultaneously provides controllability, composability, fidelity, and replicability--capabilities existing approaches struggle to achieve simultaneously. This paper introduces NetForge, a programmable substrate for bottleneck-centric data generation guided by progressive disaggregation: NetForge (i) decouples bottleneck intent from execution, (ii) separates static bottleneck attributes from dynamic congestion pressure, and (iii) disaggregates observed demand dynamics from their original trace context via Cross-Traffic Profiles (CTPs). CTPs transform passive packet traces into reusable, composable pressure signals that can be selected and transformed to specify dynamic bottleneck behavior. Our evaluation shows that NetForge satisfies the four requirements and, in an ABR case study, generates data that remains realistic, expands coverage into underrepresented regimes, and, in turn, improves model performance by up to 47% by reducing transmission-time prediction error of the Fugu model. Together, these results establish NetForge as a practical substrate for studying Internet application behavior across diverse bottleneck regimes.

cs.NI↗

arXiv2025

Leveraging Large Language Models to Contextualize Network Measurements

With the worldwide growth of remote communication and telepresence, network measurements form a cornerstone of effective performance assessment and diagnostics for Internet users. Most often, users seek for overall connection performance measurement using publicly available tools (also known as `speed tests') that provide an overview of their connection's throughput and latency. However, extracting meaningful insights from these measurements remains a challenging task for a non-technical audience. Interpreting network measurement data often requires considerable domain expertise to account not only for subtle variations of the connection stability and metrics, but even for simpler concepts such as latency under load or packet loss influence towards connection performance. In the absence of proper expertise, common misconceptions can easily arise. To address these issues, researchers should recognize the importance of making network measurements not only more comprehensive but also more accessible for wider audience without deep technical knowledge. A promising direction to achieve this goal involves leveraging recent advancements in large language models (LLMs), which have demonstrated capabilities in conducting an analysis of complex data in other fields, such as laboratory test results interpretation, news summarization, and personal assistance. In this paper, we describe an ongoing effort to apply large language models and historical data to enhance the interpretation of network measurements in real-world environments. We aim to automate the translation of low-level metric data into accessible explanations, allowing non-experts to make more informed decisions regarding network performance and reliability.

cs.NI↗

arXiv2025

The Multifractal IP Address Structure: Physical Explanation and Implications

The structure of IP addresses observed in Internet traffic plays a critical role for a wide range of networking problems of current interest. For example, modern network telemetry systems that take advantage of existing data plane technologies for line rate traffic monitoring and processing cannot afford to waste precious data plane resources on traffic that comes from "uninteresting" regions of the IP address space. However, there is currently no well-established structural model or analysis toolbox that enables a first-principles approach to the specific problem of identifying "uninteresting" regions of the address space or the myriad of other networking problems that prominently feature IP addresses. To address this key missing piece, we present in this paper a first-of-its-kind empirically validated physical explanation for why the observed IP address structure in measured Internet traffic is multifractal in nature. Our root cause analysis overcomes key limitations of mostly forgotten findings from ~20 years ago and demonstrates that the Internet processes and mechanisms responsible for how IP addresses are allocated, assigned, and used in today's Internet are consistent with and well modeled by a class of evocative mathematical models called conservative cascades. We complement this root cause analysis with the development of an improved toolbox that is tailor-made for analyzing finite and discrete sets of IP addresses and includes statistical estimators that engender high confidence in the inferences they produce. We illustrate the use of this toolbox in the context of a novel address structure anomaly detection method we designed and conclude with a discussion of a range of challenging open networking problems that are motivated or inspired by our findings.

cs.NI↗

arXiv2025

The Amazon Nova Family of Models: Technical Report and Model Card

We present Amazon Nova, a new generation of state-of-the-art foundation models that deliver frontier intelligence and industry-leading price performance. Amazon Nova Pro is a highly-capable multimodal model with the best combination of accuracy, speed, and cost for a wide range of tasks. Amazon Nova Lite is a low-cost multimodal model that is lightning fast for processing images, video, documents and text. Amazon Nova Micro is a text-only model that delivers our lowest-latency responses at very low cost. Amazon Nova Canvas is an image generation model that creates professional grade images with rich customization controls. Amazon Nova Reel is a video generation model offering high-quality outputs, customization, and motion control. Our models were built responsibly and with a commitment to customer trust, security, and reliability. We report benchmarking results for core capabilities, agentic performance, long context, functional adaptation, runtime performance, and human evaluation.

cs.AI↗

arXiv2024

NetworkGym: Reinforcement Learning Environments for Multi-Access Traffic Management in Network Simulation

Mobile devices such as smartphones, laptops, and tablets can often connect to multiple access networks (e.g., Wi-Fi, LTE, and 5G) simultaneously. Recent advancements facilitate seamless integration of these connections below the transport layer, enhancing the experience for apps that lack inherent multi-path support. This optimization hinges on dynamically determining the traffic distribution across networks for each device, a process referred to as \textit{multi-access traffic splitting}. This paper introduces \textit{NetworkGym}, a high-fidelity network environment simulator that facilitates generating multiple network traffic flows and multi-access traffic splitting. This simulator facilitates training and evaluating different RL-based solutions for the multi-access traffic splitting problem. Our initial explorations demonstrate that the majority of existing state-of-the-art offline RL algorithms (e.g. CQL) fail to outperform certain hand-crafted heuristic policies on average. This illustrates the urgent need to evaluate offline RL algorithms against a broader range of benchmarks, rather than relying solely on popular ones such as D4RL. We also propose an extension to the TD3+BC algorithm, named Pessimistic TD3 (PTD3), and demonstrate that it outperforms many state-of-the-art offline RL algorithms. PTD3's behavioral constraint mechanism, which relies on value-function pessimism, is theoretically motivated and relatively simple to implement.

cs.NI↗

arXiv2024

Redefining Proactivity for Information Seeking Dialogue

Information-Seeking Dialogue (ISD) agents aim to provide accurate responses to user queries. While proficient in directly addressing user queries, these agents, as well as LLMs in general, predominantly exhibit reactive behavior, lacking the ability to generate proactive responses that actively engage users in sustained conversations. However, existing definitions of proactive dialogue in this context do not focus on how each response actively engages the user and sustains the conversation. Hence, we present a new definition of proactivity that focuses on enhancing the `proactiveness' of each generated response via the introduction of new information related to the initial query. To this end, we construct a proactive dialogue dataset comprising 2,000 single-turn conversations, and introduce several automatic metrics to evaluate response `proactiveness' which achieved high correlation with human annotation. Additionally, we introduce two innovative Chain-of-Thought (CoT) prompts, the 3-step CoT and the 3-in-1 CoT prompts, which consistently outperform standard prompts by up to 90% in the zero-shot setting.

cs.CL↗

arXiv2024

The Efficacy of the Connect America Fund in Addressing US Internet Access Inequities

Residential fixed broadband internet access in the United States (US) has long been distributed inequitably, drawing significant attention from researchers and policymakers. This paper evaluates the efficacy of the Connect America Fund (CAF), a key policy intervention aimed at addressing disparities in US internet access. CAF subsidizes the creation of new regulated broadband monopolies in underserved areas, aiming to provide comparable internet access, in terms of price and speed, to that available in urban regions. Oversight of CAF largely relies on data self-reported by internet service providers (ISPs), which is often questionable. We use the broadband-plan querying tool (BQT) to curate a novel dataset that complements ISP-reported information with ISP-advertised broadband plan details (download speed and monthly cost) on publicly accessible websites. Specifically, we query advertised broadband plans for 687k residential addresses across 15 states, certified as served by ISPs to regulators. Our analysis reveals significant discrepancies between ISP-reported data and actual broadband availability. We find that the serviceability rate-defined as the fraction of addresses ISPs actively serve out of the total queried, weighted by the number of CAF addresses in a census block group-is only 55%, dropping to as low as 18% in some states. Additionally, the compliance rate-defined as the weighted fraction of addresses where ISPs actively serve and advertise download speeds above the FCC's 10 Mbps threshold-is only 33%. We also observe that in a subset of census blocks, CAF-funded addresses receive higher broadband speeds than their monopoly-served neighbors. These results indicate that while a few users have benefited from this multi-billion dollar program, it has largely failed to achieve its intended goal, leaving many targeted rural communities with inadequate or no broadband connectivity.

cs.NI↗