Author name cluster

Manjesh Kumar Hanawal

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

6 papers

2 author rows

ICLR Conference 2025 Conference Paper

BEEM: Boosting Performance of Early Exit DNNs using Multi-Exit Classifiers as Experts

Divya Jyoti Bajpai
Manjesh Kumar Hanawal

Early Exit (EE) techniques have emerged as a means to reduce inference latency in Deep Neural Networks (DNNs). The latency improvement and accuracy in these techniques crucially depend on the criteria used to make exit decisions. We propose a new decision criterion BEEM where exit classifiers are treated as experts and aggregate their confidence scores. The confidence scores are aggregated only if neighbouring experts are consistent in prediction as the samples pass through them, thus capturing their ensemble effect. A sample exits when the aggregated confidence value exceeds a threshold. The threshold is set using the error rates of the intermediate exits aiming to surpass the performance of conventional DNN inference. Experimental results on the COCO dataset for Image captioning and GLUE datasets for various language tasks demonstrate that our method enhances the performance of state-of-the-art EE methods, achieving improvements in speed-up by a factor $1.5\times$ to $2.1\times$. When compared to the final layer, its accuracy is comparable in harder Image Captioning and improves in the easier language tasks. The source code is available at https://github.com/Div290/BEEM1/tree/main.

Details

NeurIPS Conference 2025 Conference Paper

Beyond Greedy Exits: Improved Early Exit Decisions for Risk Control and Reliability

Divya Jyoti Bajpai
Manjesh Kumar Hanawal

Early-Exit Deep Neural Networks enable adaptive inference by allowing prediction at intermediary layers, significantly reducing computational costs and latency. Most of the early exit strategies greedily exit a sample at an intermediary layer if the confidence in class prediction exceeds a predefined threshold that is set using a static validation set. This is problematic as the model might be overconfident in a wrong class. Also, they are not robust to distribution shifts encountered in deployment, which can undermine model trustworthiness and accuracy. To address these challenges, we propose UAT that adapts the threshold for exit decisions using a Multi-Armed Bandit framework, enabling online, unsupervised adjustment of exit decisions. UAT makes decisions based on a new reward function that assesses predictive certainty and its reliability to balance computational efficiency and prediction quality while penalizing unnecessary late exits. We provide guarantees on risk achieved by UAT and validate its performance on diverse tasks spanning vision-language understanding, text generation, and classification. Our framework demonstrates consistent improvements in speedup $(1. 70-2. 10\times)$ with a minimal performance drop $(<2)$\% as compared to full model performance.

PDF Details

TMLR Journal 2024 Journal Article

Fixed Budget Best Arm Identification in Unimodal Bandits

Debamita Ghosh
Manjesh Kumar Hanawal
Nikola Zlatanov

We consider the best arm identification problem in a fixed budget stochastic multi-armed bandit in which arm means exhibit unimodal structure, i.e., there is only one local maximum. We establish that the probability of misidentifying the optimal arm within a budget of $T$ is lower bounded as $\mathcal{O}\left(\exp\left\{-T/\bar{H}\right\}\right)$, where $\bar{H}$ depends on the sub-optimality gaps of arms in the neighborhood of the optimal arm. % where $\bar{H}\leq 2\Delta^{-2}$. In contrast to the lower bound for the unstructured case, the error exponent in this bound does not depend on the number of arms $K$ and is smaller by a factor $\log K$, which captures the gain achievable by exploiting the unimodal structure. We then develop an algorithm named {\it Fixed Budget Best Arm Unimodal Bandits ( FB-BAUB)} that exploits unimodality to achieve the gain. Specifically, we show that the error probability of \algo{} is upper bounded as $\mathcal{O}\left(\log_2 K\exp\left\{-T\Delta^2\right\}\right)$, where $\Delta$ is the gap between the neighboring arms and $\bar{H}\leq 2\Delta^{-2}$. We demonstrate that \algo{} outperforms the state-of-the-art algorithms through extensive simulations. Moreover, \algo{} is parameter-free and simple to implement.

PDF Details

NeurIPS Conference 2021 Conference Paper

Stochastic Multi-Armed Bandits with Control Variates

Arun Verma
Manjesh Kumar Hanawal

This paper studies a new variant of the stochastic multi-armed bandits problem where auxiliary information about the arm rewards is available in the form of control variates. In many applications like queuing and wireless networks, the arm rewards are functions of some exogenous variables. The mean values of these variables are known a priori from historical data and can be used as control variates. Leveraging the theory of control variates, we obtain mean estimates with smaller variance and tighter confidence bounds. We develop an upper confidence bound based algorithm named UCB-CV and characterize the regret bounds in terms of the correlation between rewards and control variates when they follow a multivariate normal distribution. We also extend UCB-CV to other distributions using resampling methods like Jackknifing and Splitting. Experiments on synthetic problem instances validate performance guarantees of the proposed algorithms.

PDF Details

NeurIPS Conference 2020 Conference Paper

Online Algorithm for Unsupervised Sequential Selection with Contextual Information

Arun Verma
Manjesh Kumar Hanawal
Csaba Szepesvari
Venkatesh Saligrama

In this paper, we study Contextual Unsupervised Sequential Selection (USS), a new variant of the stochastic contextual bandits problem where the loss of an arm cannot be inferred from the observed feedback. In our setup, arms are associated with fixed costs and are ordered, forming a cascade. In each round, a context is presented, and the learner selects the arms sequentially till some depth. The total cost incurred by stopping at an arm is the sum of fixed costs of arms selected and the stochastic loss associated with the arm. The learner's goal is to learn a decision rule that maps contexts to arms with the goal of minimizing the total expected loss. The problem is challenging as we are faced with an unsupervised setting as the total loss cannot be estimated. Clearly, learning is feasible only if the optimal arm can be inferred (explicitly or implicitly) from the problem structure. We observe that learning is still possible when the problem instance satisfies the so-called 'Contextual Weak Dominance' (CWD) property. Under CWD, we propose an algorithm for the contextual USS problem and demonstrate that it has sub-linear regret. Experiments on synthetic and real datasets validate our algorithm.

PDF Details

ICML Conference 2015 Conference Paper

Cheap Bandits

Manjesh Kumar Hanawal
Venkatesh Saligrama
Michal Valko
Rémi Munos

We consider stochastic sequential learning problems where the learner can observe the average reward of several actions. Such a setting is interesting in many applications involving monitoring and surveillance, where the set of the actions to observe represent some (geographical) area. The importance of this setting is that in these applications, it is actually cheaper to observe average reward of a group of actions rather than the reward of a single action. We show that when the reward is smooth over a given graph representing the neighboring actions, we can maximize the cumulative reward of learning while minimizing the sensing cost. In this paper we propose CheapUCB, an algorithm that matches the regret guarantees of the known algorithms for this setting and at the same time guarantees a linear cost again over them. As a by-product of our analysis, we establish a Ω(\sqrt(dT)) lower bound on the cumulative regret of spectral bandits for a class of graphs with effective dimension d.

Details