Search

Your search keyword '"Sharma, Akshat"' showing total 141 results
Start Over Author "Sharma, Akshat"
141 results on '"Sharma, Akshat"'

1. MiniKV: Pushing the Limits of LLM Inference via 2-Bit Layer-Discriminative KV Cache

Author

Sharma, Akshat, Ding, Hangliang, Li, Jianping, Dani, Neel, and Zhang, Minjia

Subjects
Computer Science - Computation and Language, Computer Science - Machine Learning
Abstract

How to efficiently serve LLMs in practice has become exceptionally challenging due to their prohibitive memory and computation requirements. In this study, we investigate optimizing the KV cache, whose memory footprint poses a critical bottleneck in LLM inference, especially when dealing with long context tasks. To tackle the challenge, we introduce MiniKV, a KV cache optimization method that simultaneously preserves long context task accuracy while significantly reducing KV cache size via a novel 2-bit layer-discriminative KV cache. More importantly, we develop specialized CUDA kernels to make MiniKV compatible with FlashAttention. Experiments on a wide range of long context tasks show that MiniKV effectively achieves 86% KV cache compression ratio while recovering over 98.5% of accuracy, outperforming state-of-the-art methods while achieving excellent measured system performance improvements.

Published
2024

2. IL-TUR: Benchmark for Indian Legal Text Understanding and Reasoning

Author

Joshi, Abhinav, Paul, Shounak, Sharma, Akshat, Goyal, Pawan, Ghosh, Saptarshi, and Modi, Ashutosh

Subjects
Computer Science - Computation and Language, Computer Science - Artificial Intelligence, Computer Science - Machine Learning
Abstract

Legal systems worldwide are inundated with exponential growth in cases and documents. There is an imminent need to develop NLP and ML techniques for automatically processing and understanding legal documents to streamline the legal system. However, evaluating and comparing various NLP models designed specifically for the legal domain is challenging. This paper addresses this challenge by proposing IL-TUR: Benchmark for Indian Legal Text Understanding and Reasoning. IL-TUR contains monolingual (English, Hindi) and multi-lingual (9 Indian languages) domain-specific tasks that address different aspects of the legal system from the point of view of understanding and reasoning over Indian legal documents. We present baseline models (including LLM-based) for each task, outlining the gap between models and the ground truth. To foster further research in the legal domain, we create a leaderboard (available at: https://exploration-lab.github.io/IL-TUR/) where the research community can upload and compare legal text understanding systems., Comment: Accepted at ACL 2024 Main Conference; 40 Pages (9 Pages + References + Appendix)

Published
2024

3. Adaptive Thresholding Heuristic for KPI Anomaly Detection

Author

Isaac, Ebenezer R. H. P. and Sharma, Akshat

Subjects
Computer Science - Machine Learning, Statistics - Machine Learning
Abstract

A plethora of outlier detectors have been explored in the time series domain, however, in a business sense, not all outliers are anomalies of interest. Existing anomaly detection solutions are confined to certain outlier detectors limiting their applicability to broader anomaly detection use cases. Network KPIs (Key Performance Indicators) tend to exhibit stochastic behaviour producing statistical outliers, most of which do not adversely affect business operations. Thus, a heuristic is required to capture the business definition of an anomaly for time series KPI. This article proposes an Adaptive Thresholding Heuristic (ATH) to dynamically adjust the detection threshold based on the local properties of the data distribution and adapt to changes in time series patterns. The heuristic derives the threshold based on the expected periodicity and the observed proportion of anomalies minimizing false positives and addressing concept drift. ATH can be used in conjunction with any underlying seasonality decomposition method and an outlier detector that yields an outlier score. This method has been tested on EON1-Cell-U, a labeled KPI anomaly dataset produced by Ericsson, to validate our hypothesis. Experimental results show that ATH is computationally efficient making it scalable for near real time anomaly detection and flexible with multiple forecasters and outlier detectors.

Published
2023

4. U-CREAT: Unsupervised Case Retrieval using Events extrAcTion

Author

Joshi, Abhinav, Sharma, Akshat, Tanikella, Sai Kiran, and Modi, Ashutosh

Subjects
Computer Science - Information Retrieval, Computer Science - Artificial Intelligence, Computer Science - Computation and Language, Computer Science - Machine Learning
Abstract

The task of Prior Case Retrieval (PCR) in the legal domain is about automatically citing relevant (based on facts and precedence) prior legal cases in a given query case. To further promote research in PCR, in this paper, we propose a new large benchmark (in English) for the PCR task: IL-PCR (Indian Legal Prior Case Retrieval) corpus. Given the complex nature of case relevance and the long size of legal documents, BM25 remains a strong baseline for ranking the cited prior documents. In this work, we explore the role of events in legal case retrieval and propose an unsupervised retrieval method-based pipeline U-CREAT (Unsupervised Case Retrieval using Events Extraction). We find that the proposed unsupervised retrieval method significantly increases performance compared to BM25 and makes retrieval faster by a considerable margin, making it applicable to real-time case retrieval systems. Our proposed system is generic, we show that it generalizes across two different legal systems (Indian and Canadian), and it shows state-of-the-art performance on the benchmarks for both the legal systems (IL-PCR and COLIEE corpora)., Comment: Accepted at ACL 2023, 15 pages (12 main + 3 Appendix)

Published
2023

10. Genetically Optimized PID Controller for a Novel Corn Dryer

Author

Abhigna, Raptadu, Sharma, Akshat, Kavya, Kudumuu, Sharma, Pankaj, Salkuti, Surender Reddy, Rajesh Kumar, M., Naidu, Rani Chinnappa, Sreekeessoon, Bhamini, Kacprzyk, Janusz, Series Editor, Gomide, Fernando, Advisory Editor, Kaynak, Okyay, Advisory Editor, Liu, Derong, Advisory Editor, Pedrycz, Witold, Advisory Editor, Polycarpou, Marios M., Advisory Editor, Rudas, Imre J., Advisory Editor, Wang, Jun, Advisory Editor, Thakur, Manoj, editor, Agnihotri, Samar, editor, Rajpurohit, Bharat Singh, editor, Pant, Millie, editor, Deep, Kusum, editor, and Nagar, Atulya K., editor

Published
2023
Full Text
View/download PDF

20. The acute management of ankle fractures (Augment) study: A prospective trainee led national collaborative audit of the Boast 12 guidelines

Author

Yousuf, Umar, Senthilkumaran, Sriskandarasa, Kumar, Vish, Korde, Vinayak, Baigent, Thomas, Mahmood, Fahd, Shin, Je Song, Haines, Samuel, Dominguez, Elizabeth, Monsell, Fergal, Armbruster, Berenice Aguirrezabala, Hakim, Muhammad Aadil, Dewar, Greg, Kozhikunnath, Arun, Rai, Praveen, Adebayo, Oliver, Reddy, Rajesh Gopi, Mishra, Arya, Tadros, John, Black, Jonathon, Willmott, Henry, Robertson, Alastair, Rae, Fraser, Macleod, Calum, Macleod, Lewis, Tan, Harriet, Smith, Ben, Chuter, Graham, Slade, Henry, Javed, Omar, Adamu-Biu, Fatima, Ho, Beatrice, Williams, Mark, Stedman, Tobias, Bull, Eleanor, Abiddin, Zain, Hannah, Andrew, Fananapazir, Leila, Del Rosario-Caluya, Christine, Hassan, Abdul Rahman, Howe, Thomas, Carr, Alastair, Arnaout, Firas, Knapper, Thomas, Fleming, Tom, Everett, Samuel James, Jordan, Stevan, Fowler, Tim, Grice, John, Taylor, Alexander, Crosswell, Sebastien, Campion, Alice, Bissell, Iain, Al-Kaisi, Kahlan, Paavana, Thumri, Neal-Smith, Gregory, Bretherton, Chris, Dowlut, Naeem, Gourbault, Lysander, Goh, En Lin, Costa, Matthew, Wadkin, James, Harris, Maredudd, Gill, Ian, Siddiqui, Nashat, White, Andrew, Chawla, Amit, Dalwai, Ramesh, Dixit, Anurag, Papachristou, Athanasios, Sharma, Akshat, Thomas, Osmond, Clifton, Laura, Wong, Huan, Tweedie, Blair, Drake, Benjamin, Khan, Sameer, Townshend, David, Barnard, Emily, Ward, Alex, Madhvani, Kiran, de Boer, Henry, Clark, Matthew, Naveed, Umna, Chahal, Gurdip, Watts, Anna, Barkley, Sarah, Doorgakant, Ashtin, Blundell, Chris, Price, Michael, Harrison, Peter, Wright, Jake, Lesko, Laszlo, Osman, Khabab, Hartley, Laura, Uhiara, Okezika, Marsh, Alastair, Lloyd, Adam, Naguleswaran, Sinthuja, Khamdan, Khadija, Hartshorn, Rebecca, Sood, Abhay, Divecha, Hiren, Bryony, Peiris, Jaibaji, Monketh, McAndrew, Andrew, Kennedy, James, Yusuf, Muhammad, Rogers, Sophie, Hodhody, Ghazal, Jones, Matt, King, Andrew, Agrawal, Somen, Stephens, Christian Gray, Davies-Jones, Gareth, Pirozekova, Lucia, Vick, Oliver, Kay, Robert, Keating, John, Walsh, Nicola, Elbana, Husam, Gandavaram, Srikanth, Cash, Thomas, Jayatilaka, Lara, Rourke, Harry, Finnie, Stephanie, Studnicka, Katarzyna, Holmes, David, Caruana, Ethan, Oputa, Tobi, Bux, Mohammed, McLauchlan, George, Stirling, Scott, Hoggett, Lee, Peake, Chris, Duke, Kathryn, Rossiter, Daniel, Torku, David, Solan, Matthew, Anibueze, Chukwudubem, To, Christopher, Manafa, Chibuzo, Gyimah, Dominic Ayi, Fagg, James, Duggleby, Luke, Sgardelis, Panagiotis, Stanger, Sophie, Barton, Tristan, Al-Sukaini, Ahmad, Venkatesan, Mohanraj, McKerron, Sally, Balasubramanian, Sathishkumar, Koç, Togay, Dean, Mike, Abdelaziz, Ahmed, Toh, Eugene, Stevenson, Robert Allen, Naqvi, Ali Zain, Balai, Edward, Little, Max, Pearse, Michael, Sait, Saif, Drysdale, Henry, Lowdon, Hamish, Rahim, Sybghat, Fennelly, Joseph, Davda, Kinner, Solari, Francesca, Hill, Daniel, Hanger, Joseph, Anderson, Kerry, Jones, Robin, Dutta, Agneish, Sayani, Junaid, Yeoh, Tien, Crate, Georgina, Forder, Justin, Bikkannavar, Pratik, Erotocritou, Marios, Relwani, Jiten, Singh, Saumya, Sinha, Rachel, Stringfellow, Thomas, Eves, Joshua, Asher, Dipika, Jones, Alyssa, Ali, Karar, Rajendran, Kausala, Cadwallader, Craig, Gray, Matt, Mani, Ravi, Ebreo, Darren, Mahoney, Rachel, Latif, Javed, Baid, Roshni, Charalambides, Maria, Khatri, Khushboo, Zafeiris, Konstantinos, Charalambides, Charalambos, Taylor, Isabel, Qulaghassi, Mahdi, Onubogu, Ifeanyi Kem, Oliver, Matthew, Manning, Christopher, Williams, James, Cheyne, Natalie, Pillai, Anand, Fennelly, Joseph T., Gourbault, Lysander J., Price, Michael J., and Ward, Alex E.

Published
2021
Full Text
View/download PDF

22. Human CD4+iNKT cell adoptive immunotherapy induces anti‐tumour responses against CD1d‐negative EBV‐driven B lymphoma.

Author

Baiu, Dana C., Sharma, Akshat, Schehr, Jennifer L., Basu, Jayati, Smith, Kelsey A., Ohashi, Makoto, Johannsen, Eric C., Kenney, Shannon C., and Gumperz, Jenny E.

Subjects
*CD14 antigen, *B cell lymphoma, *IMMUNOTHERAPY, *T cells, *CELL imaging, *LYMPHOMAS
Abstract

Invariant natural killer T (iNKT) cells are a conserved population of innate T lymphocytes that are uniquely suitable as off‐the‐shelf cellular immunotherapies due to their lack of alloreactivity. Two major subpopulations of human iNKT cells have been delineated, a CD4− subset that has a TH1/cytolytic profile, and a CD4+ subset that appears polyfunctional and can produce both regulatory and immunostimulatory cytokines. Whether these two subsets differ in anti‐tumour effects is not known. Using live cell imaging, we found that CD4− iNKT cells limited growth of CD1d+ Epstein–Barr virus (EBV)‐infected B‐lymphoblastoid spheroids in vitro, whereas CD4+ iNKT cells showed little or no direct anti‐tumour activity. However, the effects of the two subsets were reversed when we tested them as adoptive immunotherapies in vivo using a xenograft model of EBV‐driven human B cell lymphoma. We found that EBV‐infected B cells down‐regulated CD1d in vivo, and administering CD4− iNKT cells had no discernable impact on tumour mass. In contrast, xenotransplanted mice bearing lymphomas showed rapid reduction in tumour mass after administering CD4+ iNKT cells. Immunotherapeutic CD4+ iNKT cells trafficked to both spleen and tumour and were associated with subsequently enhanced responses of xenotransplanted human T cells against EBV. CD4+ iNKT cells also had adjuvant‐like effects on monocyte‐derived DCs and promoted antigen‐dependent responses of human T cells in vitro. These results show that allogeneic CD4+ iNKT cellular immunotherapy leads to marked anti‐tumour activity through indirect pathways that do not require tumour cell CD1d expression and that are associated with enhanced activity of antigen‐specific T cells. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

26. Accidental neodymium-doped yttrium aluminum garnet laser capsulotomy in a phakic eye

Author

Beniwal, Abhijeet, Sharma, Akshat, Shaikh, Nawazish, and Tandon, Radhika

Subjects
Magnetic alloys, Rare earth metals, Lasers, Laser, Health
Abstract

Byline: Abhijeet. Beniwal, Akshat. Sharma, Nawazish. Shaikh, Radhika. Tandon A 33-year-old lady presented with vision loss in the right eye post neodymium-doped yttrium aluminum garnet (Nd-YAG) capsulotomy in phakic eye [...]

Published
2023

27. Myopia in retinopathy of prematurity: Boon or bane

Author

Banerjee, Mousumi, Sharma, Akshat, Azad, Shorya, and Rao, Siddhartha

Subjects
Vision disorders in children -- Case studies -- Complications and side effects, Retrolental fibroplasia -- Case studies -- Complications and side effects, Myopia -- Case studies -- Complications and side effects, Health
Abstract

Byline: Mousumi. Banerjee, Akshat. Sharma, Shorya. Azad, Siddhartha. Rao Myopic changes in the fundus have been known to have a protective role in vasculopathy. The reduced metabolic demands prevent the [...]

Published
2023

28. Analyzing technical, sentimental, and machine learning algorithms for stock market prediction.

Author

Sharma, Akshat, Lohumi, Yogesh, Gangodkar, Durgaprasad, and Goyal, Ashtha

Subjects
*PROCESS capability, *STOCK prices, *INVESTORS, *STOCKS (Finance), *RANDOM forest algorithms, *MACHINE learning
Abstract

In the financial market, the use of machine learning algorithms for predicting stock prices is particularly advantageous. Stock prices, a critical factor influencing decisions made by traders, investors, and large firms, are subject to volatility influenced by various external factors, rendering prediction a complex endeavor. Accurate predictions, however, play a pivotal role in optimizing profitability within the realm of stock trading. Machine learning algorithms, endowed with the capacity to autonomously learn and improve, are well-suited to this task when integrated with sentiment and technical indicators. They possess the capability to process vast volumes of historical data, unveiling patterns that may elude immediate human perception. This paper presents an analysis that incorporates sentimental and technical analyses and employs machine learning algorithms viz. Logistic Regression, Random Forest, and Decision Tree for the prediction of stock prices. These algorithms, when combined with technical and emotional analysis, serve to efficiently forecast market behavior. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

30. 390 Discovery of MAGE-A1-specific TCR-T cell therapy candidates to expand multiplex therapy of solid tumors

Author

Tadros, Jenny, primary, Nabilsi, Nancy, additional, Jurewicz, Mollie, additional, Sharma, Akshat, additional, Jahan, Kenneth L, additional, Gaspar, Nicolas, additional, Cirelli, Kimberly M, additional, Parsons, Teagan, additional, Khokhar, Shazad A, additional, Kamalia, Shubhangi, additional, Padmanabhan, Sveta, additional, Kiaf, Badr, additional, Nayar, Ribhu, additional, Ospina, Victor, additional, Mahopatra, Alok Das, additional, Traore, Tary, additional, Boudot, Antoine J, additional, Dukaj, Livio, additional, He, Jin, additional, Kritzer, Ryan E, additional, Cristofaro, Alexander, additional, Pavuluri, Chandan K, additional, Miga, Emily, additional, Xu, Qikai, additional, Wang, Yifan, additional, Gurer, Cagan, additional, and MacBeath, Gavin, additional

Published
2023
Full Text
View/download PDF

32. Seronegative macular coloboma with multifocal choroiditis

Author

Banerjee, Mousumi, Sharma, Akshat, Azad, Shorya, and Venkatesh, Pradeep

Subjects
Uveitis -- Case studies -- Diagnosis, Coloboma -- Case studies -- Diagnosis, Health
Abstract

Byline: Mousumi. Banerjee, Akshat. Sharma, Shorya. Azad, Pradeep. Venkatesh Macular coloboma is a rare and distinct entity thought to be a consequence of a developmental abnormality or a resolved inflammation [...]

Published
2022

36. Lenticular indentation due to vitreous incarceration following open globe injury

Author

Bhayana, Amber, Mazumder, Saima, Prasad, Priyanka, Sharma, Akshat, Khokhar, Sudarshan, and Gupta, Anita

Subjects
Health
Abstract

Byline: Amber. Bhayana, Saima. Mazumder, Priyanka. Prasad, Akshat. Sharma, Sudarshan. Khokhar, Anita. Gupta A 28-year-old otherwise healthy male was operated for corneoscleral penetrating injury repair [Figure 1]a.[1] Postoperatively, anterior hyaloid [...]

Published
2022

37. Iatrogenic lens pseudo-coloboma following trabeculectomy

Author

Bhayana, Amber, Prasad, Priyanka, Namdev, Vaibhav, Sharma, Akshat, Dada, Tanuj, and Khokhar, Sudarshan

Subjects
Medical errors, Health
Abstract

Byline: Amber. Bhayana, Priyanka. Prasad, Vaibhav. Namdev, Akshat. Sharma, Tanuj. Dada, Sudarshan. Khokhar A 1-year-old bupthalmic child[1] [Haab's striae: blue arrow, [Figure 1] operated for trabeculectomy had lens notching visible [...]

Published
2022

Catalog

Books, media, physical & digital resources
See catalog results