arXiv subjects

Rajat Verma

Publications and source records attributed to Rajat Verma.

6 recordsLinked to original sources

arXiv2026

Lost in Reordering: Structural Sensitivity of Multilingual LLMs under Semantics-Preserving Perturbations

Large Language Models (LLMs) demonstrate strong multilingual reasoning performance, yet their robustness to semantics-preserving structural variation remains underexplored, particularly for relatively free word-order languages. We investigate the structural sensitivity of multilingual LLMs using two linguistically grounded perturbation settings in Hindi and Malayalam: constrained constituent reordering and active-passive voice transformation. We introduce a benchmark dataset IndicReStruct, with two variants, GSM8K-Reordered and GSM8K-Voice, constructed from GSM8K while preserving semantic meaning. Across six state-of-the-art LLMs and multiple prompting strategies, we observe consistent and significant degradation in mathematical reasoning performance under structurally perturbed inputs. To further understand these failures, we perform qualitative error analysis and mechanistic interpretability experiments using residual-stream activation patching. Our analyses show that reasoning failures frequently arise from disruptions in entity-quantity alignment and that intermediate transformer layers contribute most strongly toward reasoning restoration. Overall, our findings suggest that current multilingual LLMs remain highly sensitive to surface syntactic realization and lack robust compositional invariance under structurally different but semantically equivalent inputs.

cs.CL↗

arXiv2025

Multilingual Tokenization through the Lens of Indian Languages: Challenges and Insights

Tokenization plays a pivotal role in NLP and is fundamental to training language models. However, existing tokenizers are often skewed towards high-resource languages, limiting their effectiveness for linguistically diverse and morphologically rich languages such as those in the Indian subcontinent. In this work, we present a comprehensive empirical study of multilingual tokenization across 17 Indic languages spanning 11 scripts and two language families. We systematically evaluate the effects of (i) widely used subword algorithms: BPE and Unigram LM, (ii) script and orthography-aware normalization, (iii) vocabulary size, and (iv) multilingual vocabulary construction strategies. We use a combination of intrinsic and extrinsic evaluations to obtain the following observations: (i) script-specific normalization improves tokenization quality, (ii) Unigram LM better preserves morphological boundaries than BPE, (iii) cluster-based vocabulary construction shows improvement in downstream tasks compared to the joint method. Our findings highlight the importance of linguistically informed design choices in multilingual tokenization and offer practical guidance for building effective tokenizers for low-resource and morphologically complex languages.

cs.CL↗

arXiv2024

Long-term forecasts of statewide travel demand patterns using large-scale mobile phone GPS data: A case study of Indiana

The growth in availability of large-scale GPS mobility data from mobile devices has the potential to aid traditional travel demand models (TDMs) such as the four-step planning model, but those processing methods are not commonly used in practice. In this study, we show the application of trip generation and trip distribution modeling using GPS data from smartphones in the state of Indiana. This involves extracting trip segments from the data and inferring the phone users' home locations, adjusting for data representativeness, and using a data-driven travel time-based cost function for the trip distribution model. The trip generation and interchange patterns in the state are modeled for 2025, 2035, and 2045. Employment sectors like industry and retail are observed to influence trip making behavior more than other sectors. The travel growth is predicted to be mostly concentrated in the suburban regions, with a small decline in the urban cores. Further, although the majority of the growth in trip flows over the years is expected to come from the corridors between the major urban centers of the state, relative interzonal trip flow growth will likely be uniformly spread throughout the state. We also validate our results with the forecasts of two travel demand models, finding a difference of 5-15% in overall trip counts. Our GPS data-based demand model will contribute towards augmenting the conventional statewide travel demand model developed by the state and regional planning agencies.

econ.GN↗

arXiv2024

Towards a generalized accessibility measure for transportation equity and efficiency

Locational measures of accessibility are widely used in urban and transportation planning to understand the impact of the transportation system on influencing people's access to places. However, there is a considerable lack of measurement standards and publicly available data. We propose a generalized measure of locational accessibility that has a comprehensible form for transportation planning analysis. This metric combines the cumulative opportunities approach with gravity-based measures and is capable of catering to multiple trip purposes, travel modes, cost thresholds, and scales of analysis. Using data from multiple publicly available datasets, this metric is computed by trip purpose and travel time threshold for all block groups in the United States, and the data is made publicly accessible. Further, case studies of three large metropolitan areas reveal substantial inefficiencies in transportation infrastructure, with the most inefficiency observed in sprawling and non-core urban areas, especially for bicycling. Subsequently, it is shown that targeted investment in facilities can contribute to a more equitable distribution of accessibility to essential shopping and service facilities. By assigning greater weights to socioeconomically disadvantaged neighborhoods, the proposed metric formally incorporates equity considerations into transportation planning, contributing to a more equitable distribution of accessibility to essential services and facilities.

econ.EM↗

arXiv2023

Comparison of home detection algorithms using smartphone GPS data

Estimation of people's home locations using location-based services data from smartphones is a common task in human mobility assessment. However, commonly used home detection algorithms (HDAs) are often arbitrary and unexamined. In this study, we review existing HDAs and examine five HDAs using eight high-quality mobile phone geolocation datasets. These include four commonly used HDAs as well as an HDA proposed in this work. To make quantitative comparisons, we propose three novel metrics to assess the quality of detected home locations and test them on eight datasets across four U.S. cities. We find that all three metrics show a consistent rank of HDAs' performances, with the proposed HDA outperforming the others. We infer that the temporal and spatial continuity of the geolocation data points matters more than the overall size of the data for accurate home detection. We also find that HDAs with high (and similar) performance metrics tend to create results with better consistency and closer to common expectations. Further, the performance deteriorates with decreasing data quality of the devices, though the patterns of relative performance persist. Finally, we show how the differences in home detection can lead to substantial differences in subsequent inferences using two case studies - (i) hurricane evacuation estimation, and (ii) correlation of mobility patterns with socioeconomic status. Our work contributes to improving the transparency of large-scale human mobility assessment applications.

cs.CY↗

arXiv2021

Mobility-based contact exposure explains the disparity of spread of COVID-19 in urban neighborhoods

The rapid early spread of COVID-19 in the U.S. was experienced very differently by different socioeconomic groups and business industries. In this study, we study aggregate mobility patterns of New York City and Chicago to identify the relationship between the amount of interpersonal contact between people in urban neighborhoods and the disparity in the growth of positive cases among these groups. We introduce an aggregate Contact Exposure Index (CEI) to measure exposure due to this interpersonal contact and combine it with social distancing metrics to show its effect on positive case growth. With the help of structural equations modeling, we find that the effect of exposure on case growth was consistently positive and that it remained consistently higher in lower-income neighborhoods, suggesting a causal path of income on case growth via contact exposure. Using the CEI, schools and restaurants are identified as high-exposure industries, and the estimation suggests that implementing specific mobility restrictions on these point-of-interest categories are most effective. This analysis can be useful in providing insights for government officials targeting specific population groups and businesses to reduce infection spread as reopening efforts continue to expand across the nation.

econ.GN↗