arXiv subjects

Suman Banerjee

Publications and source records attributed to Suman Banerjee.

At least 19 recordsLinked to original sources

arXiv2026

NPUsper: Eliminating Redundant Computation for Real-Time Whisper on Mobile NPUs

We present NPUsper, a live transcription system that makes Whisper efficient on mobile NPUs by eliminating redundant computation. To avoid the heavy padding used by prior streaming systems, NPUsper detects hallucinated tokens online from temporal patterns in decoder cross-attention, allowing each inference round to process short audio inputs with minimal carryover. For efficient mobile-NPU execution, we propose controlled unrolling, which executes autoregressive decoding as K-step chunk graphs, removing unnecessary KV-cache computation and reducing graph-dispatch overhead. NPUsper achieves up to 4.84x lower per-word latency, up to 33.2x lower time-to-first-token (TTFT), and up to 88.64% lower average power consumption compared with baselines, while maintaining comparable transcription accuracy. The code is available at https://github.com/npusper/NPUsper.

cs.SD↗

arXiv2026

Exploiting RIS Optimization Limits for Multi-User Beamforming and Signal Suppression

This paper presents a unified framework for exploiting the boundaries of reconfigurable intelligent surfaces (RIS) joint optimization in multi-user wireless systems, where a single RIS accommodates diverse objectives.We first propose an adaptive gradient-scaling mechanism that accelerates the convergence of the underlying optimization algorithm while maintaining stable performance across varying channel and system parameters. The proposed mechanism enables the solver to reach a reasonably good solution rapidly without requiring manual tuning of step sizes or algorithmic hyperparameters when system inputs change. We then propose a low-complexity beamformer recovery method tailored for single-user scenarios, which circumvents the full matrix decomposition required by traditional approaches, thereby significantly reducing computational overhead. Building on these foundations, we develop an element allocation strategy that enables user-specific prioritization through assignment of RIS subsets. This is further extended by a modular add-drop mechanism that supports partial-panel optimization in general multi-user settings. The framework is evaluated across three representative scenarios: (i) signal amplification for all users, (ii) signal suppression for all users, and (iii) selective amplification and suppression. To characterize performance limits, we derive power trade-off boundaries using scalarized joint optimization, which closely align with Monte Carlo simulations. Our unified joint optimization method consistently yield solutions near these boundaries, confirming its near-optimality. Extensive simulations under realistic channel models demonstrate that the proposed approach outperforms conventional semidefinite relaxation techniques, offering a scalable and effective RIS control strategy for cooperative and competitive multi-user environments.

eess.SP↗

arXiv2026

EMBER: Efficient Memory via Budgeted Evidence Retention for Long-Horizon Agents

Long-horizon agents can archive large histories, but future answers still incur retrieval, rereading, and context costs. When retained memory misses answer-relevant evidence, the system must return to larger portions of the raw history. We study budgeted evidence survival: before the query is known, which source evidence should be retained so that it remains recoverable and usable under a fixed retained source-evidence token budget? We instantiate this setting as Budgeted Pre-Query Retention, where memory is written during ingestion and later read without access to the full raw stream. We introduce EMBER, a learned retention policy that constructs a compact, source-backed evidence state. EMBER stores evidence capsules: verbatim source excerpts paired with retrieval keys and update metadata, preserving both grounding and read-time access. Post-query outcome feedback trains the writer to preserve evidence across the ingestion-retrieval-answer chain. On LongMemEval-RR, our LongMemEval-derived retained-evidence protocol, EMBER-14B reaches 0.3017 F1 at the 8192-token retained-evidence comparison point, compared with 0.1765 for the strongest non-EMBER budgeted baseline. Across retained source-evidence budgets, EMBER improves F1, Retain-Recall, and Read-Recall, indicating that long-horizon memory depends on retaining evidence within the budget rather than rereading larger histories.

cs.CL↗

arXiv2026

Fairness-Aware Profit Maximization using Deep Reinforcement Learning

Given a social network represented as a graph where the nodes are the users and the edges represent the social relations, and a positive integer k, how to select k nodes to maximize the influence in the network remains an active area of research. In this paper, we consider a variant of the problem in which network users are associated with two parameters: a benefit value and a cost. A fixed budget is given, and the network is partitioned into communities. The task is to select a subset of users (the seed set) within the budget so that their initial activation maximizes the earned profit, while ensuring that each community realizes at least a minimum fraction of its total benefit under a maximin fairness criterion. For any seed set, the earned benefit is defined as the sum of the benefit values of the users influenced by the seed set, and the profit is defined as the difference between the earned benefit and the total cost. Formally, we call this the Fairness-Aware Profit Maximization Problem. We propose a Deep Reinforcement Learning-based approach for solving it: we first model the problem as a Markov Decision Process and subsequently propose a Deep Q-Learning Algorithm. The proposed solution has been implemented and tested on real-world social network datasets. From the reported results, we observed that the proposed approach yields a seed set whose initial activation produces up to 10 times more profit than the baseline methods. The implementation of our methodology is available at https://github.com/PoonamSharma-PY/DRL_FPM.git.

cs.SI↗

arXiv2026

Cast a Wider Net: Coordinated Pass@K Policy Optimization for Code Reasoning

Repeated sampling with a verifier is the standard way to allocate test-time compute for code generation, with pass@$K$ as the canonical metric. Yet the standard policy class draws $K$ independent samples from a single answer distribution, so attempts often collapse onto near-duplicate reasoning paths and waste the budget on redundant rollouts. This failure is costly in competitive programming, where many problems admit multiple distinct algorithmic strategies and pass@$K$ requires only one correct attempt. We propose Coordinated Pass@$K$ Policy Optimization (CPPO), which turns pass@$K$ generation into joint exploration over strategies: a planner emits a tuple of $K{=}4$ alternative high-level methods, and a shared solver attempts one solution per method. CPPO trains this joint policy with a multiplicative planner reward, $R_{\mathrm{plan}} = J_\psi \cdot R_{\mathrm{out}}$, assigning credit only to valid strategy tuples that lead to verifier-confirmed pass@$K$ success. Across APPS, CodeContests, and LiveCodeBench-v6, CPPO improves pass@$4$ over direct sampling, planning baselines, planner-only SFT, and pass@$K$-oriented RL under the same $K{=}4$ solver-attempt budget, with statistically significant gains on six of nine model--benchmark cells. The largest single gain is $+0.16$ on Qwen3.5-9B LiveCodeBench-v6 over the strongest baseline, PKPO ($0.588 \rightarrow 0.748$; paired bootstrap, $p < 0.05$).

cs.CL↗

arXiv2026

Counting Small Balanced (p,q)-bicliques in Signed Bipartite Graphs

Two disjoint sets of entities and their relationship can be modelled as a bipartite graph. Real-life examples include drug-target interaction in biological networks, user-item relationships in e-commerce networks, etc. Motif-based analysis is essential for understanding the structure of large-scale networks, and bipartite graphs are no exception. In contrast to unsigned graphs, motif analysis in signed bipartite graphs has received limited attention. The smallest non-trivial motif in a signed bipartite graph is a balanced (2,2)-biclique, often called a balanced butterfly, which captures only local patterns and cannot reveal higher-order relationships. Bipartite motifs have been studied in the literature in the context of signed bipartite graphs, such as maximal biclique, bitruss, and so on. None of these works addresses bipartite motifs with fixed-sized vertex sets, which are often relevant in practical situations. In this work, we study the balanced (p,q)-biclique counting problem for small values of p and q. As a baseline, we first adapt and extend the state-of-the-art BCList++ algorithm for unsigned bipartite graphs to incorporate edge signs, which we call SBCList++. We then propose two efficient algorithms: BBWC, a wedge-centric approach that enforces balance constraints during enumeration, and BBVP, a vertex-based pruning approach that directly enumerates feasible vertex sets. Extensive experiments on large real-world datasets demonstrate that the vertex-based pruning algorithm, BBVP, significantly outperforms the baseline, achieving an average speedup of 636$\times$ over SBCList++ (where p=q=3).

cs.DS↗

arXiv2026

Accelerating Transformer-Based Monocular SLAM via Geometric Utility Scoring

Geometric Foundation Models (GFMs) have recently advanced monocular SLAM by providing robust, calibration-free 3D priors. However, deploying these models on dense video streams introduces significant computational redundancy. Current GFM-based SLAM systems typically rely on post hoc keyframe selection. Because of this, they must perform expensive dense geometric decoding simply to determine whether a frame contains novel geometry, resulting in late rejection and wasted computation. To mitigate this inefficiency, we propose LeanGate, a lightweight feed-forward frame-gating network. LeanGate predicts a geometric utility score to assess a frame's mapping value prior to the heavy GFM feature extraction and matching stages. As a predictive plug-and-play module, our approach bypasses over 90% of redundant frames. Evaluations on standard SLAM benchmarks demonstrate that LeanGate reduces tracking FLOPs by more than 85% and achieves a 5x end-to-end throughput speedup. Furthermore, it maintains the tracking and mapping accuracy of dense baselines. Project page: https://lean-gate.github.io/

cs.CV↗

arXiv2026

Approximation Algorithms for Budget Splitting in Multi-Channel Influence Maximization

How to utilize an allocated budget effectively for branding and promotion of a commercial house is an important problem, particularly when multiple advertising media are available. There exist multiple such media, and among them, two popular ones are billboards and social media advertisements. In this context, the question naturally arises: how should a budget be allocated to maximize total influence? Although there is significant literature on the effective use of budgets in individual advertising media, there are hardly any studies examining budget allocation across multiple advertising media. To bridge this gap, this paper introduces the \textsc{Budget Splitting Problem in Billboard and Social Network Advertisement}. We introduce the notion of \emph{interaction effect} to capture the additional influence due to triggers from multiple media of advertising. Using this notion, we propose a noble influence function $\Phi(,)$ that captures the total influence and shows that this function is non-negative, monotone, and non-bisubmodular. We introduce \emph{bi-submodularity ratio $(\gamma)$} and \emph{generalized curvature $(\alpha)$} to measure how close a function is to being bi-submodular and how far a function is from being modular, respectively. We propose the Randomized Greedy and Two-Phase Adaptive Greedy approach, where the influence function is non-bisubmodular and achieves an approximation guarantee of $\frac{1}{\alpha}\left(1-e^ {-\gamma \alpha} \right)$. We conducted several experiments using real-world datasets and observed that the proposed solution approach's budget splitting leads to a greater influence than existing approaches.

cs.DS↗

arXiv2026

Tag-specific Regret Minimization Problem in Outdoor Advertising

Recently, out-of-home advertising has become a popular marketing technique, due to its higher return on investment. E-commerce houses approach the influence provider to achieve effective advertising through their tags (advertising content), influence demand, and budgets. The influence provider's goal will be to make proper tag allocations, meet the required influence demand within the budget constraint, and minimize total regret. We formalize this as a combinatorial optimization problem and refer to it as \textsc{Tag-specific Regret Minimization in Outdoor Advertising (TRMOA)}. We show that TRMOA is NP-hard and inapproximable within a constant factor. The regret model we consider is non-monotone and non-submodular, and the simple greedy approach is ineffective. We introduce a fairness-aware greedy round-robin approach that reduces regret with balanced allocation across advertisers. To improve, we also introduce randomized greedy and local search algorithms. We have experimented with all the methodologies using real-world trajectory and billboard datasets to show the effectiveness and efficiency of the solution methodologies.

cs.DB↗

arXiv2026

Profit Maximization in Closed Social Networks

Diffusion of information, innovation, and ideas is an important phenomenon in social networks. Information propagates through the network and reaches from one person to the next. In many settings, it is meaningful to restrict diffusion so that each node can spread information to only a limited number of its neighbors rather than to all of them. Such social networks are called closed social networks. In recent years, social media platforms have emerged as an effective medium for commercial entities, where the objective is to maximize profit. In this paper, we study the Profit Maximization in Closed Social Networks (PMCSN) problem in the context of viral marketing. The input to the problem is a closed social network and two positive integers $\ell$ and $B$. The problem asks to select seed nodes within a given budget $B$; during the diffusion process, each node is restricted to choose at most $\ell$ outgoing links for information diffusion; and the objective is to maximize the profit earned by the seed set. The PMCSN problem generalizes the Influence Maximization problem, which is NP-hard. We propose two solution approaches for PMCSN: a sampling-based approximate solution and a marginal-gain-based heuristic solution. We analyze the sample complexity, running time, and space requirements of the proposed approaches. We conduct experiments on real-world, publicly available social network datasets. The results show that the seed sets and diffusion links chosen by our methods yield higher profit than baseline methods. The implementation and data are available at \texttt{https://github.com/PoonamSharma-PY/ClosedNetwork}.

cs.SI↗

arXiv2026

DeepPM: A Deep Learning-based Profit Maximization Approach in Social Networks

The problem of Profit Maximization asks to choose a limited number of influential users from a given social network such that the initial activation of these users maximizes the profit earned at the end of the diffusion process. This problem has a direct impact on viral marketing in social networks. Over the past decade, several traditional methodologies (i.e., non-learning-based, which include approximate solution, heuristic solution, etc.) have been developed, and many of them produce promising results. All these methods require the information diffusion model as input. However, it may not be realistic to consider any particular diffusion model as real-world diffusion scenarios will be much more complex and need not follow the rules for any particular diffusion model. In this paper, we propose a deep learning-based framework to solve the profit maximization problem. Our model makes a latent representation of the seed sets and is able to learn the diversified information diffusion pattern. We also design a noble objective function that can be optimized effectively using the proposed learning-based approach. The proposed model has been evaluated with the real-world datasets, and the results are reported. We compare the effectiveness of the proposed approach with many existing methods and observe that the seed set chosen by the proposed learning-based approach leads to more profit compared to existing methods. The whole implementation and the simulation code is available at: https://github.com/PoonamSharma-PY/DeepPM.

cs.SI↗

arXiv2026

Multi-core & GPU-based Balanced Butterfly Counting in Signed Bipartite Graphs

Balanced butterfly counting, corresponding to counting balanced (2, 2)-bicliques, is a fundamental primitive in the analysis of signed bipartite graphs and provides a basis for studying higher-order structural properties such as clustering coefficients and community structure. Although prior work has proposed an efficient CPU-based serial method for counting balanced (2, k)-bicliques. The computational cost of balanced butterfly counting remains a major bottleneck on large-scale graphs. In this work, we present the highly parallel implementations for balanced butterfly counting for both multicore CPUs and GPUs. The proposed multi-core algorithm (M-BBC) employs fine-grained vertex-level parallelism to accelerate wedge-based counting while eliminating the generation of unbalanced substructures. To improve scalability, we develop a GPU-based method (G-BBC) that uses a tile-based parallel approach to effectively leverage shared memory while handling large vertex sets. We then present an improved variation, G-BBC++, which integrates dynamic scheduling to mitigate workload imbalance and maximize throughput. We conduct an experimental assessment of the proposed methods across 15 real-world datasets. Experimental results exhibit that M-BBC achieves speedups of up to 71.13x (average 38.13x) over the sequential baseline BB2K. The GPU-based algorithms deliver even greater improvements, achieving up to 13,320x speedup (average 2,600x) over BB2K and outperforming M-BBC by up to 186x (average 50x). These results indicate the substantial scalability and efficiency of our parallel algorithms and establish a robust foundation for high-performance signed motif analysis on massive bipartite graphs.

cs.DC↗

arXiv2026

Profit Maximization for Viral Marketing in Online Social Networks using Two Phase Diffusion Approach

Now-a-days, Online Social Networks (OSNs) are extensively used by different commercial houses for viral marketing. The key problem that arises in this context is to choose a limited number of highly influential users as the initial adopters of a brand such that the influence regarding the brand in the network gets maximized. Deviating from this standard setting, in this paper, we study the problem where every user of the network is associated with a selection cost and a benefit value. This benefit value can be earned from the user if (s)he is influenced by the brand. A fixed amount of budget is allocated for selecting the seed users. The goal of initial adopters is to choose a set of seed users within the budget such that the profit is maximized. We propose a two phase diffusion model for this problem where the goal is to split the diffusion process into two phases, and hence, split the budget into two halves. First, we spend the first half budget to select seed users for the first phase and observe the diffusion for a few rounds and then deploy the seed users for the second phase and successively complete the diffusion process. We prove several properties of the two phase influence function. Three solution approaches have been proposed for our problem with detailed analysis and illustrative examples. We conduct a number of experiments with three real-world social network datasets. From the experiments, we observe that the two phase diffusion approach leads to more amount of profit compared to the single-phase diffusion. In particular, for most instances, this improvement is greater than 18% and reaching as high as 40% by the proposed methodologies.

cs.SI↗

arXiv2026

Trajectory-Driven Multi-Product Influence Maximization in Billboard Advertising

Billboard Advertising has emerged as an effective out-of-home advertising technique, where the goal is to select a limited number of slots and play advertisement content there, with the hope that it will be observed by many people and, effectively, a significant number of them will be influenced towards the brand. Given a trajectory and a billboard database and a positive integer $k$, how can we select $k$ highly influential slots to maximize influence? In this paper, we study a variant of this problem where a commercial house wants to make a promotion of multiple products, and there is an influence demand for each product. We have studied two variants of the problem. In the first variant, our goal is to select $k$ slots such that the respective influence demand of each product is satisfied. In the other variant of the problem, we are given with $\ell$ integers $k_1,k_2, \ldots, k_{\ell}$, the goal here is to search for $\ell$ many set of slots $S_1, S_2, \ldots, S_{\ell}$ such that for all $i \in [\ell]$, $|S_{i}| \leq k_i$ and for all $i \neq j$, $S_i \cap S_j=\emptyset$ and the influence demand of each of the products gets satisfied. We model the first variant of the problem as a multi-submodular cover problem and the second variant as its generalization. To solve the common-slot variant, we formulate the problem as a multi-submodular cover problem and design a bi-criteria approximation algorithm based on the continuous greedy framework and randomized rounding. For the disjoint-slot variant, we proposed a sampling-based approximation approach along with an efficient primal-dual greedy algorithm that enforces disjointness naturally. Extensive experiments with real-world trajectory and billboard datasets highlight the effectiveness and efficiency of the proposed solution approaches.

cs.DB↗

arXiv2026

Fairness Driven Multi-Agent Path Finding Problem

The Multi-Agent Path Finding (MAPF) problem aims at finding non-conflicting paths for multiple agents from their respective sources to destinations. This problem arises in multiple real-life situations, including robot motion planning and airspace assignment for unmanned aerial vehicle movement. The problem is computationally expensive, and adding to it, the agents are rational and can misreport their private information. In this paper, we study both variants of the problem under the realm of fairness. For the non-rational agents, we propose a heuristic solution for this problem. Considering the agents are rational, we develop a mechanism and demonstrate that it is a dominant strategy, incentive compatible, and individually rational. We employ various solution methodologies to highlight the effectiveness and efficiency of the proposed solution approaches.

cs.MA↗

arXiv2025

ByteLoom: Weaving Geometry-Consistent Human-Object Interactions through Progressive Curriculum Learning

Human-object interaction (HOI) video generation has garnered increasing attention due to its promising applications in digital humans, e-commerce, advertising, and robotics imitation learning. However, existing methods face two critical limitations: (1) a lack of effective mechanisms to inject multi-view information of the object into the model, leading to poor cross-view consistency, and (2) heavy reliance on fine-grained hand mesh annotations for modeling interaction occlusions. To address these challenges, we introduce ByteLoom, a Diffusion Transformer (DiT)-based framework that generates realistic HOI videos with geometrically consistent object illustration, using simplified human conditioning and 3D object inputs. We first propose an RCM-cache mechanism that leverages Relative Coordinate Maps (RCM) as a universal representation to maintain object's geometry consistency and precisely control 6-DoF object transformations in the meantime. To compensate HOI dataset scarcity and leverage existing datasets, we further design a training curriculum that enhances model capabilities in a progressive style and relaxes the demand of hand mesh. Extensive experiments demonstrate that our method faithfully preserves human identity and the object's multi-view geometry, while maintaining smooth motion and object manipulation.

cs.CV↗

arXiv2025

A Reverse Reachable Set Based Approach for Motif Oriented Profit maximization in Social Networks

Profit Maximization is one of the key objectives for social media marketing, where the task is to choose a limited number of highly influential nodes such that their initial activation leads to maximum profit. In this paper, we introduce a variant of the Profit Maximization Problem where we consider that instead of nodes, benefits are assigned to some of the motifs of the graph, and these benefit values can be earned once a given threshold count of nodes from the motifs is influenced. The goal here is to choose a limited number of nodes for initial activation called seed nodes such that the motif-oriented profit gets maximized. Formally, we call our problem the Motif Oriented Profit Maximization Problem. We show that the problem is NP-hard to solve optimally. We propose a Reverse Reachable Set-based framework to solve our problem. The proposed methodology broadly divides into three steps: KPT Estimation and RR Set generation, Seed Set Selection, and Motif Oriented Profit Estimation. The proposed methodology has been analyzed to understand its time and space requirements. It has been implemented with real-world social network datasets, and the results are reported. We observe that the seed set selected by the proposed solution approaches leads to more profit compared to the seed sets selected by the existing methods. The whole implementation and data are available at: https://github.com/PoonamSharma-PY/MotifProfit.

cs.SI↗

arXiv2025

Privacy-Aware Sharing of Raw Spatial Sensor Data for Cooperative Perception

Cooperative perception between vehicles is poised to offer robust and reliable scene understanding. Recently, we are witnessing experimental systems research building testbeds that share raw spatial sensor data for cooperative perception. While there has been a marked improvement in accuracies and is the natural way forward, we take a moment to consider the problems with such an approach for eventual adoption by automakers. In this paper, we first argue that new forms of privacy concerns arise and discourage stakeholders to share raw sensor data. Next, we present SHARP, a research framework to minimize privacy leakage and drive stakeholders towards the ambitious goal of raw data based cooperative perception. Finally, we discuss open questions for networked systems, mobile computing, perception researchers, industry and government in realizing our proposed framework.

cs.NI↗