Asmit Kumar Singh

Asmit Kumar Singh, Haozhe Wang, Laxmi Naga Santosh Attaluri et al.

Large language models (LLMs) now sit in the critical path of search, assistance, and agentic workflows, making semantic caching essential for reducing inference cost and latency. Production deployments typically use a tiered static-dynamic design: a static cache of curated, offline vetted responses mined from logs, backed by a dynamic cache populated online. In practice, both tiers are commonly governed by a single embedding similarity threshold, which induces a hard tradeoff: conservative thresholds miss safe reuse opportunities, while aggressive thresholds risk serving semantically incorrect responses. We introduce \textbf{Krites}, an asynchronous, LLM-judged caching policy that expands static coverage without changing serving decisions. On the critical path, Krites behaves exactly like a standard static threshold policy. When the nearest static neighbor of the prompt falls just below the static threshold, Krites asynchronously invokes an LLM judge to verify whether the static response is acceptable for the new prompt. Approved matches are promoted into the dynamic cache, allowing future repeats and paraphrases to reuse curated static answers and expanding static reach over time. In trace-driven simulations on conversational and search workloads, Krites increases the fraction of requests served with curated static answers (direct static hits plus verified promotions) by up to $\textbf{3.9}$ times for conversational traffic and search-style queries relative to tuned baselines, with unchanged critical path latency.

Khalid Shaikh, Asmit Kumar Singh, Rebecca Christopher Dsouza et al.

Large language model (LLM) inference is increasingly bottlenecked by the Key-Value (KV) cache, yet the fine-grained structure of attention-head activations remains poorly understood. We show that pretrained Transformers exhibit a pervasive inter-head linear structure: for a given token, the Query, Key, and Value (QKV) vectors of an attention head can often be reconstructed as a linear combination of a small number of peer heads, typically within the same layer. Across Llama-3.1-8B, Falcon3-10B, OLMo-2-7B, and Qwen3-32B, just 2-5 reference heads recover many target heads with high fidelity (e.g., mean R^2 approx 0.76 for Keys on C4 with five references, and frequently R^2 > 0.85 on GSM8K). This predictability is learned rather than architectural: it is largely absent at random initialization, rises rapidly during pretraining as we track through OLMo-2 checkpoints, and is supported by a theoretical lower bound showing high mean-squared error for linear prediction at initialization. We further connect this emergence to increasing intra-layer alignment of Key projection subspaces. Finally, we exploit this redundancy for efficiency by caching only reference-head KV states and reconstructing the remaining heads on the fly via lightweight linear maps, achieving 2x KV-cache reduction with model-dependent accuracy trade-offs (4.5-5.5 percentage point average drop on Falcon3-10B and Qwen3-32B across five benchmarks, and larger drops on Llama-3.1-8B), and we find that reconstructing Keys is substantially less harmful than reconstructing Values.

Asmit Kumar Singh, Paras Mehan, Divyanshu Sharma et al.

Wearing masks is a useful protection method against COVID-19, which has caused widespread economic and social impact worldwide. Across the globe, governments have put mandates for the use of face masks, which have received both positive and negative reaction. Online social media provides an exciting platform to study the use of masks and analyze underlying mask-wearing patterns. In this article, we analyze 2.04 million social media images for six US cities. An increase in masks worn in images is seen as the COVID-19 cases rose, particularly when their respective states imposed strict regulations. We also found a decrease in the posting of group pictures as stay-at-home laws were put into place. Furthermore, mask compliance in the Black Lives Matter protest was analyzed, eliciting that 40% of the people in group photos wore masks, and 45% of them wore the masks with a fit score of greater than 80%. We introduce two new datasets, VAriety MAsks - Classification (VAMA-C) and VAriety MAsks - Segmentation (VAMA-S), for mask detection and mask fit analysis tasks, respectively. For the analysis, we create two frameworks, face mask detector (for classifying masked and unmasked faces) and mask fit analyzer (a semantic segmentation based model to calculate a mask-fit score). The face mask detector achieved a classification accuracy of 98%, and the semantic segmentation model for the mask fit analyzer achieved an Intersection Over Union (IOU) score of 98%. We conclude that such a framework can be used to evaluate the effectiveness of such public health strategies using social media platforms in times of pandemic.

Saurabh Gupta, Asmit Kumar Singh, Arun Balaji Buduru et al.

Hashtags in online social media have become a way for users to build communities around topics, promote opinions, and categorize messages. In the political context, hashtags on Twitter are used by users to campaign for their parties, spread news, or to get followers and get a general idea by following a discussion built around a hashtag. In the past, researchers have studied certain types and specific properties of hashtags by utilizing a lot of data collected around hashtags. In this paper, we perform a large-scale empirical analysis of elections using only the hashtags shared on Twitter during the 2019 Lok Sabha elections in India. We study the trends and events unfolded on the ground, the latent topics to uncover representative hashtags and semantic similarity to relate hashtags with the election outcomes. We collect over 24 million hashtags to perform extensive experiments. First, we find the trending hashtags to cross-reference them with the tweets in our dataset to list down notable events. Second, we use Latent Dirichlet Allocation to find topic patterns in the dataset. In the end, we use skip-gram word embedding model to find semantically similar hashtags. We propose popularity and an influence metric to predict election outcomes using just the hashtags. Empirical results show that influence is a good measure to predict the election outcome.