arXiv subjects

Gaurav Mishra

Publications and source records attributed to Gaurav Mishra.

15 recordsLinked to original sources

arXiv2026

PSSI-MaxST: An Efficient Pixel-Segment Similarity Index Using Intensity and Smoothness Features for Maximum Spanning Tree Based Segmentation

Interactive graph-based segmentation methods partition an image into foreground and background regions with the aid of user inputs. However, existing approaches often suffer from high computational costs, sensitivity to user interactions, and degraded performance when the foreground and background share similar color distributions. A key factor influencing segmentation performance is the similarity measure used for assigning edge weights in the graph. To address these challenges, we propose a novel Pixel Segment Similarity Index (PSSI), which leverages the harmonic mean of inter-channel similarities by incorporating both pixel intensity and spatial smoothness features. The harmonic mean effectively penalizes dissimilarities in any individual channel, enhancing robustness. The computational complexity of PSSI is $\mathcal{O}(B)$, where $B$ denotes the number of histogram bins. Our segmentation framework begins with low-level segmentation using MeanShift, which effectively captures color, texture, and segment shape. Based on the resulting pixel segments, we construct a pixel-segment graph with edge weights determined by PSSI. For partitioning, we employ the Maximum Spanning Tree (MaxST), which captures strongly connected local neighborhoods beneficial for precise segmentation. The integration of the proposed PSSI, MeanShift, and MaxST allows our method to jointly capture color similarity, smoothness, texture, shape, and strong local connectivity. Experimental evaluations on the GrabCut and Images250 datasets demonstrate that our method consistently outperforms current graph-based interactive segmentation methods such as AMOE, OneCut, and SSNCut in terms of segmentation quality, as measured by Jaccard Index (IoU), $F_1$ score, execution time and Mean Error (ME). Code is publicly available at: https://github.com/KaustubhShejole/PSSI-MaxST.

cs.CV↗

arXiv2025

Investigating Thematic Patterns and User Preferences in LLM Interactions using BERTopic

This study applies BERTopic, a transformer-based topic modeling technique, to the lmsys-chat-1m dataset, a multilingual conversational corpus built from head-to-head evaluations of large language models (LLMs). Each user prompt is paired with two anonymized LLM responses and a human preference label, used to assess user evaluation of competing model outputs. The main objective is uncovering thematic patterns in these conversations and examining their relation to user preferences, particularly if certain LLMs are consistently preferred within specific topics. A robust preprocessing pipeline was designed for multilingual variation, balancing dialogue turns, and cleaning noisy or redacted data. BERTopic extracted over 29 coherent topics including artificial intelligence, programming, ethics, and cloud infrastructure. We analysed relationships between topics and model preferences to identify trends in model-topic alignment. Visualization techniques included inter-topic distance maps, topic probability distributions, and model-versus-topic matrices. Our findings inform domain-specific fine-tuning and optimization strategies for improving real-world LLM performance and user satisfaction.

cs.LG↗

arXiv2025

Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities

In this report, we introduce the Gemini 2.X model family: Gemini 2.5 Pro and Gemini 2.5 Flash, as well as our earlier Gemini 2.0 Flash and Flash-Lite models. Gemini 2.5 Pro is our most capable model yet, achieving SoTA performance on frontier coding and reasoning benchmarks. In addition to its incredible coding and reasoning skills, Gemini 2.5 Pro is a thinking model that excels at multimodal understanding and it is now able to process up to 3 hours of video content. Its unique combination of long context, multimodal and reasoning capabilities can be combined to unlock new agentic workflows. Gemini 2.5 Flash provides excellent reasoning abilities at a fraction of the compute and latency requirements and Gemini 2.0 Flash and Flash-Lite provide high performance at low latency and cost. Taken together, the Gemini 2.X model generation spans the full Pareto frontier of model capability vs cost, allowing users to explore the boundaries of what is possible with complex agentic problem solving.

cs.CL↗

arXiv2024

Nirjas: An open source framework for extracting metadata from the source code

Metadata and comments are critical elements of any software development process. In this paper, we explain how metadata and comments in source code can play an essential role in comprehending software. We introduce a Python-based open-source framework, Nirjas, which helps in extracting this metadata in a structured manner. Various syntaxes, types, and widely accepted conventions exist for adding comments in source files of different programming languages. Edge cases can create noise in extraction, for which we use Regex to accurately retrieve metadata. Non-Regex methods can give results but often miss accuracy and noise separation. Nirjas also separates different types of comments, source code, and provides details about those comments, such as line number, file name, language used, total SLOC, etc. Nirjas is a standalone Python framework/library and can be easily installed via source or pip (the Python package installer). Nirjas was initially created as part of a Google Summer of Code project and is currently developed and maintained under the FOSSology organization.

cs.SE↗

arXiv2024

Training Language Models on the Knowledge Graph: Insights on Hallucinations and Their Detectability

While many capabilities of language models (LMs) improve with increased training budget, the influence of scale on hallucinations is not yet fully understood. Hallucinations come in many forms, and there is no universally accepted definition. We thus focus on studying only those hallucinations where a correct answer appears verbatim in the training set. To fully control the training data content, we construct a knowledge graph (KG)-based dataset, and use it to train a set of increasingly large LMs. We find that for a fixed dataset, larger and longer-trained LMs hallucinate less. However, hallucinating on $\leq5$% of the training data requires an order of magnitude larger model, and thus an order of magnitude more compute, than Hoffmann et al. (2022) reported was optimal. Given this costliness, we study how hallucination detectors depend on scale. While we see detector size improves performance on fixed LM's outputs, we find an inverse relationship between the scale of the LM and the detectability of its hallucinations.

cs.CL↗

arXiv2023

Frontier Language Models are not Robust to Adversarial Arithmetic, or "What do I need to say so you agree 2+2=5?

We introduce and study the problem of adversarial arithmetic, which provides a simple yet challenging testbed for language model alignment. This problem is comprised of arithmetic questions posed in natural language, with an arbitrary adversarial string inserted before the question is complete. Even in the simple setting of 1-digit addition problems, it is easy to find adversarial prompts that make all tested models (including PaLM2, GPT4, Claude2) misbehave, and even to steer models to a particular wrong answer. We additionally provide a simple algorithm for finding successful attacks by querying those same models, which we name "prompt inversion rejection sampling" (PIRS). We finally show that models can be partially hardened against these attacks via reinforcement learning and via agentic constitutional loops. However, we were not able to make a language model fully robust against adversarial arithmetic attacks.

cs.CL↗

arXiv2023

PaLM 2 Technical Report

We introduce PaLM 2, a new state-of-the-art language model that has better multilingual and reasoning capabilities and is more compute-efficient than its predecessor PaLM. PaLM 2 is a Transformer-based model trained using a mixture of objectives. Through extensive evaluations on English and multilingual language, and reasoning tasks, we demonstrate that PaLM 2 has significantly improved quality on downstream tasks across different model sizes, while simultaneously exhibiting faster and more efficient inference compared to PaLM. This improved efficiency enables broader deployment while also allowing the model to respond faster, for a more natural pace of interaction. PaLM 2 demonstrates robust reasoning capabilities exemplified by large improvements over PaLM on BIG-Bench and other reasoning tasks. PaLM 2 exhibits stable performance on a suite of responsible AI evaluations, and enables inference-time control over toxicity without additional overhead or impact on other capabilities. Overall, PaLM 2 achieves state-of-the-art performance across a diverse set of tasks and capabilities. When discussing the PaLM 2 family, it is important to distinguish between pre-trained models (of various sizes), fine-tuned variants of these models, and the user-facing products that use these models. In particular, user-facing products typically include additional pre- and post-processing steps. Additionally, the underlying models may evolve over time. Therefore, one should not expect the performance of user-facing products to exactly match the results reported in this report.

cs.CL↗

arXiv2022

Scaling Instruction-Finetuned Language Models

Finetuning language models on a collection of datasets phrased as instructions has been shown to improve model performance and generalization to unseen tasks. In this paper we explore instruction finetuning with a particular focus on (1) scaling the number of tasks, (2) scaling the model size, and (3) finetuning on chain-of-thought data. We find that instruction finetuning with the above aspects dramatically improves performance on a variety of model classes (PaLM, T5, U-PaLM), prompting setups (zero-shot, few-shot, CoT), and evaluation benchmarks (MMLU, BBH, TyDiQA, MGSM, open-ended generation). For instance, Flan-PaLM 540B instruction-finetuned on 1.8K tasks outperforms PALM 540B by a large margin (+9.4% on average). Flan-PaLM 540B achieves state-of-the-art performance on several benchmarks, such as 75.2% on five-shot MMLU. We also publicly release Flan-T5 checkpoints, which achieve strong few-shot performance even compared to much larger models, such as PaLM 62B. Overall, instruction finetuning is a general method for improving the performance and usability of pretrained language models.

cs.LG↗

arXiv2022

PaLI: A Jointly-Scaled Multilingual Language-Image Model

Effective scaling and a flexible task interface enable large language models to excel at many tasks. We present PaLI (Pathways Language and Image model), a model that extends this approach to the joint modeling of language and vision. PaLI generates text based on visual and textual inputs, and with this interface performs many vision, language, and multimodal tasks, in many languages. To train PaLI, we make use of large pre-trained encoder-decoder language models and Vision Transformers (ViTs). This allows us to capitalize on their existing capabilities and leverage the substantial cost of training them. We find that joint scaling of the vision and language components is important. Since existing Transformers for language are much larger than their vision counterparts, we train a large, 4-billion parameter ViT (ViT-e) to quantify the benefits from even larger-capacity vision models. To train PaLI, we create a large multilingual mix of pretraining tasks, based on a new image-text training set containing 10B images and texts in over 100 languages. PaLI achieves state-of-the-art in multiple vision and language tasks (such as captioning, visual question-answering, scene-text understanding), while retaining a simple, modular, and scalable design.

cs.CV↗

arXiv2022

Detailed investigation on x-ray emission from laser driven high-Z foils in a wide intensity range : role of conversion layer and reemission zone

Detailed radiation hydrodynamic simulations are carried out to investigate x-ray emission process in four high-Z planar targets namely, tungsten (W), gold (Au), lead (Pb) and uranium (U) irradiated by 1 ns, 351 nm flat top laser pulses. A thorough zoning analysis is performed for all laser driven high-Z foils over a wide intensity range of $10^{12}-10^{15} W/cm^{2}$ with appropriately chosen photon energy range and recombination parameter. The resulting variation of conversion efficiency over the full intensity range exhibits an optimum for all materials which is explained by considering the characteristic emission contributions from two different regions of laser irradiated plasma, namely, conversion layer and remission zone. A new generalized single scaling relation based upon smooth broken power law is proposed for conversion efficiency variation along with the separate determination ($\eta_{S}$, $\eta_{M}$) in soft and hard/M-band x-ray regions. It has been observed that $\eta_{S}$ for Pb and W always lies in between that for Au and U for intensities smaller than $\sim 3\times 10^{13} W/cm^{2}$. On further increase in intensity, $\eta_{S}$ is observed to be maximum for Au and U whereas it is minimum for W. Significant contribution to M-band conversion efficiencies is observed in all elements for intensities higher than $\sim 2\times 10^{13} W/cm^{2}$ with maximum and minimum values attained by W and U, respectively. The results are explained by considering the contributions from the emission coefficients of all materials in both conversion layer and reemission zone up to corresponding photon cut-off energies at different laser intensities.

physics.plasm-ph↗

arXiv2022

PaLM: Scaling Language Modeling with Pathways

Large language models have been shown to achieve remarkable performance across a variety of natural language tasks using few-shot learning, which drastically reduces the number of task-specific training examples needed to adapt the model to a particular application. To further our understanding of the impact of scale on few-shot learning, we trained a 540-billion parameter, densely activated, Transformer language model, which we call Pathways Language Model PaLM. We trained PaLM on 6144 TPU v4 chips using Pathways, a new ML system which enables highly efficient training across multiple TPU Pods. We demonstrate continued benefits of scaling by achieving state-of-the-art few-shot learning results on hundreds of language understanding and generation benchmarks. On a number of these tasks, PaLM 540B achieves breakthrough performance, outperforming the finetuned state-of-the-art on a suite of multi-step reasoning tasks, and outperforming average human performance on the recently released BIG-bench benchmark. A significant number of BIG-bench tasks showed discontinuous improvements from model scale, meaning that performance steeply increased as we scaled to our largest model. PaLM also has strong capabilities in multilingual tasks and source code generation, which we demonstrate on a wide array of benchmarks. We additionally provide a comprehensive analysis on bias and toxicity, and study the extent of training data memorization with respect to model scale. Finally, we discuss the ethical considerations related to large language models and discuss potential mitigation strategies.

cs.CL↗

arXiv2022

Scaling Up Models and Data with $\texttt{t5x}$ and $\texttt{seqio}$

Recent neural network-based language models have benefited greatly from scaling up the size of training datasets and the number of parameters in the models themselves. Scaling can be complicated due to various factors including the need to distribute computation on supercomputer clusters (e.g., TPUs), prevent bottlenecks when infeeding data, and ensure reproducible results. In this work, we present two software libraries that ease these issues: $\texttt{t5x}$ simplifies the process of building and training large language models at scale while maintaining ease of use, and $\texttt{seqio}$ provides a task-based API for simple creation of fast and reproducible training data and evaluation pipelines. These open-source libraries have been used to train models with hundreds of billions of parameters on datasets with multiple terabytes of training data. Along with the libraries, we release configurations and instructions for T5-like encoder-decoder models as well as GPT-like decoder-only architectures. $\texttt{t5x}$ and $\texttt{seqio}$ are open source and available at https://github.com/google-research/t5x and https://github.com/google/seqio, respectively.

cs.LG↗

arXiv2021

Effect of soft and hard x-rays on shock propagation, preheating and ablation characteristics in pure and doped Be ablators

In this paper, we analyze the performance of pure and doped Be ablators used for Inertial Confinement Fusion (ICF) pellets in terms of shock velocity, shock breakout temperature, preheat temperature and mass ablation rate through radiation hydrodynamic (RHD) simulations. For this study, we apply a constant radiation profile (drive temperatures varying from 120 - 200 eV) consisting of a low frequency Planck spectrum (soft x-rays) and a high frequency Gaussian spectrum (hard x-rays, commonly termed as "M-band") on a planar foil of the ablator. The fraction of energy density in the hard x-ray spectrum ($\alpha$) has been varied from 0 to 0.25. The predominant effect of hard x-rays is to preheat the ablator ahead of the shock front. Steady rise in preheat temperature and shock breakout temperature is observed on increasing the fraction of hard x-rays. Preheating can be mitigated by doping Be with a mid-Z element Cu as its opacity is much higher in the high frequency region. On doping Be with 1\% Cu, the shock velocities decrease slightly compared to pure Be. However, higher shock velocities are observed on increasing the fraction of M-band. We observe significant decrease in shock breakout and maximum preheat temperature in doped Be foil. Steady rise in these temperatures is observed on increasing $\alpha$. We have proposed new scaling relations for shock velocity, shock breakout temperature, maximum preheat temperature and mass ablation rate with the radiation temperature and the fraction of M-band energy density in both pure and doped Be ablators. In terms of ablator performance, Cu doped Be ablator is found to be superior to pure Be. Though doping significantly reduces preheating, the mass ablation rates are nearly unaffected.

physics.plasm-ph↗

arXiv2020

Explainable Disease Classification via weakly-supervised segmentation

Deep learning based approaches to Computer Aided Diagnosis (CAD) typically pose the problem as an image classification (Normal or Abnormal) problem. These systems achieve high to very high accuracy in specific disease detection for which they are trained but lack in terms of an explanation for the provided decision/classification result. The activation maps which correspond to decisions do not correlate well with regions of interest for specific diseases. This paper examines this problem and proposes an approach which mimics the clinical practice of looking for an evidence prior to diagnosis. A CAD model is learnt using a mixed set of information: class labels for the entire training set of images plus a rough localisation of suspect regions as an extra input for a smaller subset of training images for guiding the learning. The proposed approach is illustrated with detection of diabetic macular edema (DME) from OCT slices. Results of testing on on a large public dataset show that with just a third of images with roughly segmented fluid filled regions, the classification accuracy is on par with state of the art methods while providing a good explanation in the form of anatomically accurate heatmap /region of interest. The proposed solution is then adapted to Breast Cancer detection from mammographic images. Good evaluation results on public datasets underscores the generalisability of the proposed solution.

cs.CV↗

arXiv2016

Effects of cluster size and spatial laser intensity distribution on fusion neutron generation by laser driven Deuterium clusters

A three dimensional molecular dynamic code is used to study the generation of fusion neutrons from Coulomb explosion of Deuterium clusters driven by intense near infra-red (NIR) laser ($\lambda=800nm$) of femtosecond pulse duration ($\tau=50 fs$) under beam-target interaction scheme. We have considered various clusters of average sizes ($\langle R_0\rangle$=80,140,200\AA) which are irradiated by a laser of peak spatial-temporal intensity of 1$\times10^{18} W/cm^{2}$. The effects of cluster size and spatial laser intensity distribution on ion energies due to the Coulomb explosion of the cluster are included by convolution of single cluster single intensity ion energy distribution function (IEDF) over a range of cluster sizes and laser intensities. The final convoluted IEDF gets broadened on both lower and higher energy sides due to this procedure. Furthermore, the neutron yield which takes into account the convoluted IEDF, also gets modified by a factor of $\sim$2 compared to the case when convolution effects are ignored.

physics.plasm-ph↗