arXiv subjects

Rohan Gupta

Publications and source records attributed to Rohan Gupta.

12 recordsLinked to original sources

arXiv2025

RobotFleet: An Open-Source Framework for Centralized Multi-Robot Task Planning

Coordinating heterogeneous robot fleets to achieve multiple goals is challenging in multi-robot systems. We introduce an open-source and extensible framework for centralized multi-robot task planning and scheduling that leverages LLMs to enable fleets of heterogeneous robots to accomplish multiple tasks. RobotFleet provides abstractions for planning, scheduling, and execution across robots deployed as containerized services to simplify fleet scaling and management. The framework maintains a shared declarative world state and two-way communication for task execution and replanning. By modularizing each layer of the autonomy stack and using LLMs for open-world reasoning, RobotFleet lowers the barrier to building scalable multi-robot systems. The code can be found here: https://github.com/therohangupta/robot-fleet.

cs.RO↗

arXiv2025

PULSE-A Mission Overview: Optical Communications for Undergraduate Students

Recent advances in the size, weight, and power (SWaP) requirements for space-based sensing have dramatically increased the demand for high-bandwidth downlink. However, high data rate RF transceivers still pose significant SWaP and cost restrictions, especially for university-class CubeSat missions. Optical communication may provide a solution to this challenge, enabling data transmission with order-of-magnitude rate increases over RF while being both secure and SWaP-efficient. The Polarization-modUlated Laser Satellite Experiment (PULSE-A) is a University of Chicago mission to demonstrate optical downlink at a data rate of up to 10 Mbps using circular polarization shift keying (CPolSK). PULSE-A comprises a <1.5U Optical Transmission Terminal, 3U CubeSat Bus, Optical Ground Station (OGS) employing an amateur telescope, and RF Ground Station (RFGS), all of which are being designed and integrated by a team of over 60 undergraduate students. The mission objective is threefold: (1) to provide hands-on educational experiences for undergraduate students, (2) to make hardware for optical communication systems more accessible via open-source design, and (3) to explore the viability and potential advantages of using CPolSK for optical downlink. In this work, we present an overview of the mission, and we describe the PULSE-A Team's learning-oriented approach to program management and engineering. We especially emphasize the importance of student leadership in PULSE-A's development process and the resulting benefits for the University of Chicago community. We also highlight takeaways from the experience of founding and operating an undergraduate student-led CubeSat program.

physics.ed-ph↗

arXiv2025

RL-Obfuscation: Can Language Models Learn to Evade Latent-Space Monitors?

Latent-space monitors aim to detect undesirable behaviours in Large Language Models by leveraging their internal representations rather than relying solely on black-box outputs. These methods have shown promise in identifying behaviours such as deception and unsafe completions. However, these monitors may themselves become training signals, for example, by using problematic samples found in deployment to retrain models. This raises an important question: can models learn to evade such monitors? To evaluate this capability, we introduce RL-Obfuscation, in which LLMs are finetuned via reinforcement learning to evade latent-space monitors while maintaining their blackbox behaviour. We apply RL-Obfuscation to Language Models ranging from 7B to 14B parameters and evaluate their Evasion Success Rate against a suite of monitors. We find that token-level monitors are highly vulnerable to this attack while more holistic monitors, such as max-pooling or attention-based probes, remain robust. Moreover, for these vulnerable monitors, models trained to evade a single static monitor can generalise to evade other unseen monitors. We also find that the models can be trained to conditionally bypass latent-space monitors on only certain inputs. Finally, we study how the models bypass these monitors and find that the model can learn to repurpose tokens to have different internal representations.

cs.LG↗

arXiv2025

MIB: A Mechanistic Interpretability Benchmark

How can we know whether new mechanistic interpretability methods achieve real improvements? In pursuit of lasting evaluation standards, we propose MIB, a Mechanistic Interpretability Benchmark, with two tracks spanning four tasks and five models. MIB favors methods that precisely and concisely recover relevant causal pathways or causal variables in neural language models. The circuit localization track compares methods that locate the model components - and connections between them - most important for performing a task (e.g., attribution patching or information flow routes). The causal variable localization track compares methods that featurize a hidden vector, e.g., sparse autoencoders (SAEs) or distributed alignment search (DAS), and align those features to a task-relevant causal variable. Using MIB, we find that attribution and mask optimization methods perform best on circuit localization. For causal variable localization, we find that the supervised DAS method performs best, while SAE features are not better than neurons, i.e., non-featurized hidden vectors. These findings illustrate that MIB enables meaningful comparisons, and increases our confidence that there has been real progress in the field.

cs.LG↗

arXiv2025

Four sub-Earth planets orbiting Barnard's Star from MAROON-X and ESPRESSO

Barnard's Star is an old, single M dwarf star that comprises the second-closest extrasolar system. It has a long history of claimed planet detections from both radial velocities and astrometry. However, none of these claimed detections have so far withstood further scrutiny. Continuing this story, extreme precision radial velocity (EPRV) measurements from the ESPRESSO instrument have recently been used to identify four new sub-Earth-mass planet candidates around Barnard's Star. We present here 112 radial velocities of Barnard's Star from the MAROON-X instrument that were obtained independently to search for planets around this compelling object. The data have a typical precision of 30\,cm\,s$^{-1}$ and are contemporaneous with the published ESPRESSO measurements (2021 -- 2023). The MAROON-X data on their own confirm planet b ($P$\,=\,3.154\,d) and planet candidates c and d ($P$\,=\,4.124\,d and 2.340\,d, respectively). Furthermore, adding the MAROON-X data to the ESPRESSO data strengthens the evidence for planet candidate e ($P$\,=\,6.739\,d), thus leading to its confirmation. The signals from all four planets are $<$50\,cm\,s$^{-1}$, the minimum masses of the planets range from 0.19 to 0.34\,$M_{\oplus}$, and the system is among the most compact known among late M dwarfs hosting low-mass planets. The current data rule out planets with masses $>0.57\,M_{\oplus}$ (with a $99\%$ detection probability) in Barnard Star's habitable zone ($P$\,=\,10 -- 42\,d).

astro-ph.EP↗

arXiv2025

Efficient Evaluation of Multi-Task Robot Policies With Active Experiment Selection

Evaluating learned robot control policies to determine their physical task-level capabilities costs experimenter time and effort. The growing number of policies and tasks exacerbates this issue. It is impractical to test every policy on every task multiple times; each trial requires a manual environment reset, and each task change involves re-arranging objects or even changing robots. Naively selecting a random subset of tasks and policies to evaluate is a high-cost solution with unreliable, incomplete results. In this work, we formulate robot evaluation as an active testing problem. We propose to model the distribution of robot performance across all tasks and policies as we sequentially execute experiments. Tasks often share similarities that can reveal potential relationships in policy behavior, and we show that natural language is a useful prior in modeling these relationships between tasks. We then leverage this formulation to reduce the experimenter effort by using a cost-aware expected information gain heuristic to efficiently select informative trials. Our framework accommodates both continuous and discrete performance outcomes. We conduct experiments on existing evaluation data from real robots and simulations. By prioritizing informative trials, our framework reduces the cost of calculating evaluation metrics for robot policies across many tasks.

cs.RO↗

arXiv2025

FragmentNet: Adaptive Graph Fragmentation for Graph-to-Sequence Molecular Representation Learning

Molecular representation learning methods typically tokenize molecules as individual atoms or use rigid, rule-based fragment decompositions, limiting their ability to capture meaningful chemical substructure context. We introduce FragmentNet, a graph-to-sequence model built around a novel adaptive, learned tokenizer that decomposes molecular graphs into chemically valid fragments of adjustable granularity, complemented by chemically aware spatial positional encodings that preserve molecular topology in the resulting sequence. Extending masked pre-training strategies from natural language processing to the molecular domain, we mask and reconstruct molecules at the level of chemically meaningful fragments rather than individual atoms. Evaluating across multiple property prediction benchmarks, we find that pre-training at fragment granularity leads to improved downstream performance on the majority of tasks, demonstrating that tokenization granularity is an important design choice for molecular representation learning.

cs.LG↗

arXiv2024

InterpBench: Semi-Synthetic Transformers for Evaluating Mechanistic Interpretability Techniques

Mechanistic interpretability methods aim to identify the algorithm a neural network implements, but it is difficult to validate such methods when the true algorithm is unknown. This work presents InterpBench, a collection of semi-synthetic yet realistic transformers with known circuits for evaluating these techniques. We train simple neural networks using a stricter version of Interchange Intervention Training (IIT) which we call Strict IIT (SIIT). Like the original, SIIT trains neural networks by aligning their internal computation with a desired high-level causal model, but it also prevents non-circuit nodes from affecting the model's output. We evaluate SIIT on sparse transformers produced by the Tracr tool and find that SIIT models maintain Tracr's original circuit while being more realistic. SIIT can also train transformers with larger circuits, like Indirect Object Identification (IOI). Finally, we use our benchmark to evaluate existing circuit discovery techniques.

cs.LG↗

arXiv2024

Contrast Sets for Evaluating Language-Guided Robot Policies

Robot evaluations in language-guided, real world settings are time-consuming and often sample only a small space of potential instructions across complex scenes. In this work, we introduce contrast sets for robotics as an approach to make small, but specific, perturbations to otherwise independent, identically distributed (i.i.d.) test instances. We investigate the relationship between experimenter effort to carry out an evaluation and the resulting estimated test performance as well as the insights that can be drawn from performance on perturbed instances. We use the relative performance change of different contrast set perturbations to characterize policies at reduced experimenter effort in both a simulated manipulation task and a physical robot vision-and-language navigation task. We encourage the use of contrast set evaluations as a more informative alternative to small scale, i.i.d. demonstrations on physical robots, and as a scalable alternative to industry-scale real world evaluations.

cs.RO↗

arXiv2023

Optical Script Identification for multi-lingual Indic-script

Script identification and text recognition are some of the major domains in the application of Artificial Intelligence. In this era of digitalization, the use of digital note-taking has become a common practice. Still, conventional methods of using pen and paper is a prominent way of writing. This leads to the classification of scripts based on the method they are obtained. A survey on the current methodologies and state-of-art methods used for processing and identification would prove beneficial for researchers. The aim of this article is to discuss the advancement in the techniques for script pre-processing and text recognition. In India there are twelve prominent Indic scripts, unlike the English language, these scripts have layers of characteristics. Complex characteristics such as similarity in text shape make them difficult to recognize and analyze, thus this requires advance preprocessing methods for their accurate recognition. A sincere attempt is made in this survey to provide a comparison between all algorithms. We hope that this survey would provide insight to a researcher working not only on Indic scripts but also other languages.

cs.AI↗

arXiv2021

WASSA@IITK at WASSA 2021: Multi-task Learning and Transformer Finetuning for Emotion Classification and Empathy Prediction

This paper describes our contribution to the WASSA 2021 shared task on Empathy Prediction and Emotion Classification. The broad goal of this task was to model an empathy score, a distress score and the overall level of emotion of an essay written in response to a newspaper article associated with harm to someone. We have used the ELECTRA model abundantly and also advanced deep learning approaches like multi-task learning. Additionally, we also leveraged standard machine learning techniques like ensembling. Our system achieves a Pearson Correlation Coefficient of 0.533 on sub-task I and a macro F1 score of 0.5528 on sub-task II. We ranked 1st in Emotion Classification sub-task and 3rd in Empathy Prediction sub-task

cs.CL↗

arXiv2021

MCL@IITK at SemEval-2021 Task 2: Multilingual and Cross-lingual Word-in-Context Disambiguation using Augmented Data, Signals, and Transformers

In this work, we present our approach for solving the SemEval 2021 Task 2: Multilingual and Cross-lingual Word-in-Context Disambiguation (MCL-WiC). The task is a sentence pair classification problem where the goal is to detect whether a given word common to both the sentences evokes the same meaning. We submit systems for both the settings - Multilingual (the pair's sentences belong to the same language) and Cross-Lingual (the pair's sentences belong to different languages). The training data is provided only in English. Consequently, we employ cross-lingual transfer techniques. Our approach employs fine-tuning pre-trained transformer-based language models, like ELECTRA and ALBERT, for the English task and XLM-R for all other tasks. To improve these systems' performance, we propose adding a signal to the word to be disambiguated and augmenting our data by sentence pair reversal. We further augment the dataset provided to us with WiC, XL-WiC and SemCor 3.0. Using ensembles, we achieve strong performance in the Multilingual task, placing first in the EN-EN and FR-FR sub-tasks. For the Cross-Lingual setting, we employed translate-test methods and a zero-shot method, using our multilingual models, with the latter performing slightly better.

cs.CL↗