We want a multi-robot team to complete complex tasks in minimum time where the locations of task-relevant objects are not known. Effective task completion requires reasoning over long horizons about the likely locations of task-relevant objects, how individual actions contribute to overall progress, and how to coordinate team efforts. Planning in this setting is extremely challenging: even when task-relevant information is partially known, coordinating which robot performs which action and when is difficult, and uncertainty introduces a multiplicity of possible outcomes for each action, which further complicates long-horizon decision-making and coordination. To address this, we propose a multi-robot planning abstraction that integrates learning to estimate uncertain aspects of the environment with model-based planning for long-horizon coordination. We demonstrate the efficient multi-stage task planning of our approach for 1, 2, and 3 robot teams over competitive baselines in large ProcTHOR household environments. Additionally, we demonstrate the effectiveness of our approach with a team of two LoCoBot mobile robots in real household settings.
@inproceedings{khanal2026multirobot,title={Multi-Robot Learning-Informed Task Planning Under Uncertainty},author={Khanal, Abhish and Paudel, Abhishek and Pham, Hung and Stein, Gregory J.},booktitle={IEEE International Conference on Robotics and Automation (ICRA)},year={2026},}
arXiv
Effective Task Planning with Missing Objects using Learning-Informed Object Search
Arnob, Raihan I.,
Merlin, Max,
Paudel, Abhishek,
Hedegaard, Benned,
Konidaris, George,
and Stein, Gregory J.
Task planning for mobile robots often assumes full environment knowledge and so popular approaches, like planning via the PDDL, cannot plan when the locations of task-critical objects are unknown. Recent learning-driven object search approaches are effective, but operate as standalone tools and so are not straightforwardly incorporated into full task planners, which must additionally determine both what objects are necessary and when in the plan they should be sought out. To address this limitation, we develop a planning framework centered around novel model-based LIOS actions: each a policy that aims to find and retrieve a single object. High-level planning treats LIOS actions as deterministic and so—informed by model-based calculations of the expected cost of each—generates plans that interleave search and execution for effective, sound, and complete learning-informed task planning despite uncertainty. Our work effectively reasons about uncertainty while maintaining compatibility with existing full-knowledge solvers. In simulated ProcTHOR homes and in the real world, our approach outperforms non-learned and learned baselines on tasks including retrieval and meal prep.
@article{arnob2026effective,title={Effective Task Planning with Missing Objects using Learning-Informed Object Search},author={Arnob, Raihan I. and Merlin, Max and Paudel, Abhishek and Hedegaard, Benned and Konidaris, George and Stein, Gregory J.},journal={arXiv preprint arXiv:2602.11468},year={2026},doi={10.48550/ARXIV.2602.11468},publisher={arXiv},}
2025
ICRA DC
Introspection for Long-Horizon Robot Planning under Uncertainty
The next generation of household service robots must not only perform day-to-day tasks effectively, but also learn and improve over time by continuously evaluating their own performance. My research focuses on introspection as a technique that enables a robot to self-assess its behavior while performing long-horizon tasks in environments that may not be fully known. We show that such introspection enables improved performance in tasks such as navigation in unknown environments, deployment-time learning and adaptation, and LLM-informed object search.
@inproceedings{paudel2025introspection,title={Introspection for Long-Horizon Robot Planning under Uncertainty},author={Paudel, Abhishek},booktitle={ICRA 2025 Doctoral Consortium},year={2025},}
ICRA FMNS
Deployment-time Selection of Prompts for LLM-informed Object Search in Partially-Known Environments
Paudel, Abhishek,
and Stein, Gregory J.
In ICRA 2025 Workshop on Foundation Models and Neuro-Symbolic AI for Robotics
2025
We present an approach for deployment-time selection of best-performing prompts and LLMs for LLM-informed object search in partially-known environments. Leveraging recent progress in both LLM-informed model-based planning and deployment-time behavior selection, we enable fast bandit-like selection of best-performing prompts and LLMs and demonstrate improved deployment-time performance in object search tasks. Experiments in simulated ProcTHOR household environments show that our bandit-like selection approach results in 6.1% lower average cost and 40.6% lower average cumulative regret over baseline UCB bandit selection.
@inproceedings{paudel2025deploymenttime,title={Deployment-time Selection of Prompts for {LLM}-informed Object Search in Partially-Known Environments},author={Paudel, Abhishek and Stein, Gregory J.},booktitle={ICRA 2025 Workshop on Foundation Models and Neuro-Symbolic AI for Robotics},year={2025},url={https://openreview.net/forum?id=TrecISByKi},}
2024
CoRL
Multi-Strategy Deployment-Time Learning and Adaptation for Navigation under Uncertainty
Paudel, Abhishek,
Xiao, Xuesu,
and Stein, Gregory J.
We present an approach for performant point-goal navigation in unfamiliar partially-mapped environments. When deployed, our robot runs multiple strategies for deployment-time learning and visual domain adaptation in parallel and quickly selects the best-performing among them. Choosing between policies as they are learned or adapted between navigation trials requires continually updating estimates of their performance as they evolve. Leveraging recent work in model-based learning-informed planning under uncertainty, we determine lower bounds on the would-be performance of newly-updated policies on old trials without needing to re-deploy them. This information constrains and accelerates bandit-like policy selection, affording quick selection of the best-performing strategy shortly after it would start to yield good performance. We validate the effectiveness of our approach in simulated maze-like environments, showing improved navigation cost and cumulative regret versus existing baselines.
@inproceedings{paudel2024multistrategy,title={Multi-Strategy Deployment-Time Learning and Adaptation for Navigation under Uncertainty},author={Paudel, Abhishek and Xiao, Xuesu and Stein, Gregory J.},booktitle={Conference on Robot Learning (CoRL)},year={2024},url={https://openreview.net/forum?id=Isp19rFFV4},}
CoRL LEAP
Enhancing Object Search by Augmenting Planning with Predictions from Large Language Models
Hossain, Shahriar,
Paudel, Abhishek,
and Stein, Gregory J.
In CoRL Workshop on Learning Effective Abstractions for Planning (LEAP)
2024
We enhance object search in unknown environments by integrating a Large Language Model (LLM) with model-based planning to quickly and reliably locate an object of interest. The LLM is prompted to produce predictions about the likelihood of finding the object of interest used to define a model for planning that the robot then uses to determine its search policy, affording both good performance due to the integration of learning and reliability due to the reliance on classical planning to find the object. From our findings on 200 random maps on the ProcTHOR dataset, our proposed LLM-informed planner, utilizing GPT-4o predictions, achieves a cost reduction of 27.1% and 30.3% compared to the standard baseline and Myopic LLM-informed baseline, respectively.
@inproceedings{hossain2024enhancing,title={Enhancing Object Search by Augmenting Planning with Predictions from Large Language Models},author={Hossain, Shahriar and Paudel, Abhishek and Stein, Gregory J.},booktitle={CoRL Workshop on Learning Effective Abstractions for Planning (LEAP)},year={2024},url={https://openreview.net/forum?id=qQPp4MZbpb},}
arXiv
Using Contextual Information for Sentence-level Morpheme Segmentation
Recent advancements in morpheme segmentation primarily emphasize word-level segmentation, often neglecting the contextual relevance within the sentence. In this study, we redefine the morpheme segmentation task as a sequence-to-sequence problem, treating the entire sentence as input rather than isolating individual words. Our findings reveal that the multilingual model consistently exhibits superior performance compared to monolingual counterparts. While our model did not surpass the performance of the current state-of-the-art, it demonstrated comparable efficacy with high-resource languages while revealing limitations in low-resource language scenarios.
@article{bhandari2024using,title={Using Contextual Information for Sentence-level Morpheme Segmentation},author={Bhandari, Prabin and Paudel, Abhishek},year={2024},doi={10.48550/ARXIV.2403.15436},publisher={arXiv},}
2023
IEEE/RSJ IROS
Data-Efficient Policy Selection for Navigation in Partial Maps via Subgoal-Based Abstraction
Paudel, Abhishek,
and Stein, Gregory J.
In IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)
2023
We present a novel approach for fast and reliable policy selection for navigation in partial maps. Leveraging the recent learning-augmented model-based Learning over Subgoals Planning (LSP) abstraction to plan, our robot reuses data collected during navigation to evaluate how well other alternative policies could have performed via a procedure we call offline alt-policy replay. Costs from offline alt-policy replay constrain policy selection among the LSP-based policies during deployment, allowing for improvements in convergence speed, cumulative regret and average navigation cost. With only limited prior knowledge about the nature of unseen environments, we achieve at least 67% and as much as 96% improvements on cumulative regret over the baseline bandit approach in our experiments in simulated maze and office-like environments.
@inproceedings{paudel2023dataefficient,title={Data-Efficient Policy Selection for Navigation in Partial Maps via Subgoal-Based Abstraction},author={Paudel, Abhishek and Stein, Gregory J.},booktitle={IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},year={2023},doi={10.1109/IROS55552.2023.10342047},url={https://ieeexplore.ieee.org/document/10342047},organization={IEEE},}
2022
arXiv
Motion Primitives based Path Planning with Rapidly-exploring Random Tree
We present an approach that generates kinodynamically feasible paths for robots using Rapidly-exploring Random Tree (RRT). We leverage motion primitives as a way to capture the dynamics of the robot and use these motion primitives to build branches of the tree with RRT. Since every branch is built using the robot’s motion primitives that doesn’t lead to collision with obstacles, the resulting path is guaranteed to satisfy the robot’s kinodynamic constraints and thus be feasible for navigation without any post-processing on the generated trajectory. We demonstrate the effectiveness of our approach in simulated 2D environments using simple robot models with a variety of motion primitives.
@misc{paudel2022motion,title={Motion Primitives based Path Planning with Rapidly-exploring Random Tree},author={Paudel, Abhishek},year={2022},doi={10.48550/ARXIV.2210.15784},publisher={arXiv},}
arXiv
Learning for Robot Decision Making under Distribution Shift: A Survey
With the recent advances in the field of deep learning, learning-based methods are widely being implemented in various robotic systems that help robots understand their environment and make informed decisions to achieve a wide variety of tasks or goals. However, learning-based methods have repeatedly been shown to have poor generalization when they are presented with inputs that are different from those during training leading to the problem of distribution shift. Any robotic system that employs learning-based methods is prone to distribution shift which might lead the agents to make decisions that lead to degraded performance or even catastrophic failure. In this paper, we discuss various techniques that have been proposed in the literature to aid or improve decision making under distribution shift for robotic systems. We present a taxonomy of existing literature and present a survey of existing approaches in the area based on this taxonomy. Finally, we also identify a few open problems in the area that could serve as future directions for research.
@misc{paudel2022learning,title={Learning for Robot Decision Making under Distribution Shift: A Survey},author={Paudel, Abhishek},year={2022},doi={10.48550/ARXIV.2203.07558},publisher={arXiv},}
2021
arXiv
Room Classification on Floor Plan Graphs using Graph Neural Networks
Paudel, Abhishek,
Dhakal, Roshan,
and Bhattarai, Sakshat
We present our approach to improve room classification task on floor plan maps of buildings by representing floor plans as undirected graphs and leveraging graph neural networks to predict the room categories. Rooms in the floor plans are represented as nodes in the graph with edges representing their adjacency in the map. We experiment with House-GAN dataset that consists of floor plan maps in vector format and train multilayer perceptron and graph neural networks. Our results show that graph neural networks, specifically GraphSAGE and Topology Adaptive GCN were able to achieve accuracy of 80% and 81% respectively outperforming baseline multilayer perceptron by more than 15% margin.
@misc{paudel2021room,title={Room Classification on Floor Plan Graphs using Graph Neural Networks},author={Paudel, Abhishek and Dhakal, Roshan and Bhattarai, Sakshat},year={2021},doi={10.48550/ARXIV.2108.05947},publisher={arXiv},}
arXiv
Sophisticated Students in Boston Mechanism and Gale-Shapley Algorithm for School Choice Problem
We present our experimental results of simulating the school choice problem which deals with the assignment of students to schools based on each group’s complete preference list for the other group using two algorithms: Boston mechanism and student-proposing Gale-Shapley algorithm. We compare the effects of sophisticated students altering their preference lists with regards to these two algorithms. Our simulation results show that sophisticated students can benefit more in Boston mechanism compared to Gale-Shapley algorithm based on multiple evaluation metrics.
@misc{paudel2021sophisticated,title={Sophisticated Students in Boston Mechanism and Gale-Shapley Algorithm for School Choice Problem},author={Paudel, Abhishek},year={2021},doi={10.48550/ARXIV.2108.05951},publisher={arXiv},}
2018
IEEE ICCCS
Using Personality Traits Information from Social Media for Music Recommendation
Paudel, Abhishek,
Bajracharya, Brihat Ratna,
Ghimire, Miran,
Bhattarai, Nabin,
and Baral, Daya Sagar
In IEEE International Conference on Computing, Communication and Security (ICCCS)
2018
Music is an integral part of our life. People listen to music everyday as per their taste and mood. With the advancement and increase in volume of digital content, the choice for people to listen to diverse type of music has also increased significantly. Thus, the necessity of delivering the most suited music to the listeners has been an interesting field of research in computer science. One of the important measures to deliver the best music to listeners could be their personality traits. In order to determine the personality traits of a person, social media like Facebook can be a useful platform where people express their views on different matters, share their opinions and thoughts. This paper first describes the use of Naive Bayes classifier to determine the standard Big Five Personality Traits of a person based on their status updates on Facebook profile using basic natural language processing techniques, and then proceeds to present the use of thus obtained information about personality traits to enhance the widely implemented user-to-user collaborative filtering techniques for music recommendation.
@inproceedings{paudel2018using,author={Paudel, Abhishek and Bajracharya, Brihat Ratna and Ghimire, Miran and Bhattarai, Nabin and Baral, Daya Sagar},booktitle={IEEE International Conference on Computing, Communication and Security (ICCCS)},title={Using Personality Traits Information from Social Media for Music Recommendation},year={2018},pages={116-121},doi={10.1109/CCCS.2018.8586831},url={https://ieeexplore.ieee.org/document/8586831},}