Pranav Deshpande

Mattia J. Villani, Pranav Deshpande, Akshay Seshadri et al.

Large Language Models (LLMs) often generate factually incorrect outputs, commonly termed hallucinations, that undermine trust and limit deployment in high-stakes settings. Existing hallucination detection methods typically require multiple forward passes, or access to model internals. In this work, we provide theoretical background and empirical evidence that the distribution of token-level entropies, beyond the mean captured by perplexity or length-normalised entropy, serves as a fingerprint of hallucination, with distributional shape and tail behaviour carrying independent signal. We formalize hallucination detection as a statistical hypothesis test and propose the Calibrated Entropy Score (CES), a lightweight algorithm requiring only a single forward pass and black-box access to token logits. CES combines the mean signal with the maximum signal of the generated entropy through a calibrated reference CDF, producing scores that are directly comparable across models and tasks. We establish finite-sample calibration guarantees via a novel random-length Dvoretzky--Kiefer--Wolfowitz inequality, and also prove that CES detects hallucinations with probability converging to one exponentially fast in the generation length. Across eight QA benchmarks and ten generator models spanning open-source and API access models, CES achieves the highest detection performance among all single-pass black-box methods while providing formal error guarantees that existing heuristics lack. Remarkably, CES is statistically indistinguishable from multi-sample methods that require far greater computational cost, closing the gap between lightweight and expensive detection and making it suitable for real-time, large-scale deployment.

Anuj Apte, Pranav Deshpande, Niraj Kumar et al.

Standard neural network training relies on learning-rate schedules tied to a fixed horizon, leading to strong path dependence and costly re-tuning as data availability changes. Schedule-Free (SF) methods address this by removing explicit schedules, yet SF-AdamW, the current state-of-the-art anytime optimizer, consistently underperforms well-tuned AdamW baselines. We propose SF-NorMuon, a schedule-free spectral optimizer that closes this gap: with a single hyperparameter configuration, SF-NorMuon matches or exceeds tuned AdamW on 125M and 772M parameter language models across $1$--$8\times$ Chinchilla horizons. On the theoretical side, we prove a stationarity guarantee for schedule-free spectral dynamics and identify weight decay at the fast iterate as essential for long-horizon stability. SF-NorMuon enables practitioners to obtain high-quality checkpoints at any point during training without committing to a horizon in advance. By closing the performance gap with tuned baselines, SF-NorMuon makes horizon-free optimization more practical, taking a step towards truly open-ended, continual learning.

Mengyu Chen, Pranav Deshpande, Runqing Yang et al.

Digital humans are lifelike virtual agents capable of natural conversation and are increasingly deployed in domains like retail and finance. However, most current digital humans operate in isolation from their surroundings and lack contextual awareness beyond the dialogue itself. We address this limitation by integrating ambient intelligence (AmI) - i.e., environmental sensors, IoT data, and contextual modeling - with digital human systems. This integration enables situational awareness of the user's environment, anticipatory and proactive assistance, seamless cross-device interactions, and personalized long-term user support. We present a conceptual framework defining key roles that AmI can play in shaping digital human behavior, a design space highlighting dimensions such as proactivity levels and privacy strategies, and application-driven patterns with case studies in financial and retail services. We also discuss an architecture for ambient-enabled digital humans and provide guidelines for responsible design regarding privacy and data governance. Together, our work positions ambient intelligent digital humans as a new class of interactive agents powered by AI that respond not only to users' queries but also to the context and situations in which the interaction occurs.

Feiyu Lu, Mengyu Chen, Hsiang Hsu et al.

Mixed Reality (MR) could assist users' tasks by continuously integrating virtual content with their view of the physical environment. However, where and how to place these content to best support the users has been a challenging problem due to the dynamic nature of MR experiences. In contrast to prior work that investigates optimization-based methods, we are exploring how reinforcement learning (RL) could assist with continuous 3D content placement that is aware of users' poses and their surrounding environments. Through an initial exploration and preliminary evaluation, our results demonstrate the potential of RL to position content that maximizes the reward for users on the go. We further identify future directions for research that could harness the power of RL for personalized and optimized UI and content placement in MR.

Cindy Xu, Mengyu Chen, Pranav Deshpande et al.

In this paper, we introduce a novel system designed to enhance customer service in the financial and retail sectors through a context-aware 3D virtual agent, utilizing Mixed Reality (MR) and Vision Language Models (VLMs). Our approach focuses on enabling data-driven and empathetic interactions that ensure customer satisfaction by introducing situational awareness of the physical location, personalized interactions based on customer profiles, and rigorous privacy and security standards. We discuss our design considerations critical for deployment in real-world customer service environments, addressing challenges in user data management and sensitive information handling. We also outline the system architecture and key features unique to banking and retail environments. Our work demonstrates the potential of integrating MR and VLMs in service industries, offering practical insights in customer service delivery while maintaining high standards of security and personalization.

Tyler Chen, Archan Ray, Akshay Seshadri et al.

The $k$-means algorithm (Lloyd's algorithm) is a widely used method for clustering unlabeled data. A key bottleneck of the $k$-means algorithm is that each iteration requires time linear in the number of data points, which can be expensive in big data applications. This was improved in recent works proposing quantum and quantum-inspired classical algorithms to approximate the $k$-means algorithm locally, in time depending only logarithmically on the number of data points (along with data dependent parameters) [q-means: A quantum algorithm for unsupervised machine learning, Kerenidis, Landman, Luongo, and Prakash, NeurIPS 2019; Do you know what $q$-means?, Cornelissen, Doriguello, Luongo, Tang, QTML 2025]. In this work, we describe a simple randomized mini-batch $k$-means algorithm and a quantum algorithm inspired by the classical algorithm. We demonstrate that the worst case guarantees of these algorithms can significantly improve upon the bounds for algorithms in prior work. Our improvements are due to a careful use of uniform sampling, which preserves certain symmetries of the $k$-means problem that are not preserved in previous algorithms that use data norm-based sampling.

Javier Lopez-Piqueres, Pranav Deshpande, Archan Ray et al.

We present MetaTT, a Tensor Train (TT) adapter framework for fine-tuning of pre-trained transformers. MetaTT enables flexible and parameter-efficient model adaptation by using a single shared TT to factorize transformer sub-modules. This factorization indexes key structural dimensions, including layer and matrix type, and can optionally incorporate heads and tasks. This design allows MetaTT's parameter count to scale with the sum, rather than the product, of the modes, resulting in a substantially more compact adapter. Our benchmarks compare MetaTT with LoRA along with recent state-of-the-art matrix and tensor decomposition based fine-tuning methods. We observe that when tested on single-task standard language modeling benchmarks, MetaTT achieves competitive parameter efficiency to accuracy tradeoff. We further demonstrate that MetaTT performs competitively when compared to state-of-the-art methods on multi-task learning. Finally, we leverage the TT-ansatz to design a rank adaptive optimizer inspired by the DMRG method from many-body physics. Our results demonstrate that integrating this approach with AdamW enhances optimization performance for a specified target rank.

Tyler Chen, Akshay Seshadri, Mattia J. Villani et al.

Shapley values have emerged as a critical tool for explaining which features impact the decisions made by machine learning models. However, computing exact Shapley values is difficult, generally requiring an exponential (in the feature dimension) number of model evaluations. To address this, many model-agnostic randomized estimators have been developed, the most influential and widely used being the KernelSHAP method (Lundberg & Lee, 2017). While related estimators such as unbiased KernelSHAP (Covert & Lee, 2021) and LeverageSHAP (Musco & Witter, 2025) are known to satisfy theoretical guarantees, bounds for KernelSHAP have remained elusive. We describe a broad and unified framework that encompasses KernelSHAP and related estimators constructed using both with and without replacement sampling strategies. We then prove strong non-asymptotic theoretical guarantees that apply to all estimators from our framework. This provides, to the best of our knowledge, the first theoretical guarantees for KernelSHAP and sheds further light on tradeoffs between existing estimators. Through comprehensive benchmarking on small and medium dimensional datasets for Decision-Tree models, we validate our approach against exact Shapley values, consistently achieving low mean squared error with modest sample sizes. Furthermore, we make specific implementation improvements to enable scalability of our methods to high-dimensional datasets. Our methods, tested on datasets such MNIST and CIFAR10, provide consistently better results compared to the KernelSHAP library.