Showing 1–2 of 2 results for author: Ragupathi, S
-
Exploring Pair-Wise NMT for Indian Languages
Authors:
Kartheek Akella,
Sai Himal Allu,
Sridhar Suresh Ragupathi,
Aman Singhal,
Zeeshan Khan,
Vinay P. Namboodiri,
C V Jawahar
Abstract:
In this paper, we address the task of improving pair-wise machine translation for specific low resource Indian languages. Multilingual NMT models have demonstrated a reasonable amount of effectiveness on resource-poor languages. In this work, we show that the performance of these models can be significantly improved upon by using back-translation through a filtered back-translation process and sub…
▽ More
In this paper, we address the task of improving pair-wise machine translation for specific low resource Indian languages. Multilingual NMT models have demonstrated a reasonable amount of effectiveness on resource-poor languages. In this work, we show that the performance of these models can be significantly improved upon by using back-translation through a filtered back-translation process and subsequent fine-tuning on the limited pair-wise language corpora. The analysis in this paper suggests that this method can significantly improve a multilingual model's performance over its baseline, yielding state-of-the-art results for various Indian languages.
△ Less
Submitted 10 December, 2020;
originally announced December 2020.
-
Guarding a Polygon Without Losing Touch
Authors:
Barath Ashok,
John Augustine,
Aditya Mehekare,
Sridhar Ragupathi,
Srikkanth Ramachandran,
Suman Sourav
Abstract:
We study the classical Art Gallery Problem first proposed by Klee in 1973 from a mobile multi-agents perspective. Specifically, we require an optimally small number of agents (also called guards) to navigate and position themselves in the interior of an unknown simple polygon with $n$ vertices such that the collective view of all the agents covers the polygon.
We consider the visibly connected s…
▽ More
We study the classical Art Gallery Problem first proposed by Klee in 1973 from a mobile multi-agents perspective. Specifically, we require an optimally small number of agents (also called guards) to navigate and position themselves in the interior of an unknown simple polygon with $n$ vertices such that the collective view of all the agents covers the polygon.
We consider the visibly connected setting wherein agents must remain connected through line of sight links -- a requirement particularly relevant to multi-agent systems. We first provide a centralized algorithm for the visibly connected setting that runs in time $O(n)$, which is of course optimal. We then provide algorithms for two different distributed settings. In the first setting, agents can only perceive relative proximity (i.e., can tell which of a pair of objects is closer) whereas they can perceive exact distances in the second setting. Our distributed algorithms work despite agents having no prior knowledge of the polygon. Furthermore, we provide lower bounds to show that our distributed algorithms are near-optimal.
Our visibly connected guarding ensures that (i) the guards form a connected network and (ii) the polygon is fully guarded. Consequently, this guarding provides the distributed infrastructure to execute any geometric algorithm for this polygon.
△ Less
Submitted 12 May, 2020;
originally announced May 2020.
