<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Abhijit Chunduru</title><link>https://abhijit4-debug.github.io/</link><atom:link href="https://abhijit4-debug.github.io/index.xml" rel="self" type="application/rss+xml"/><description>Abhijit Chunduru</description><generator>Hugo Blox Builder (https://hugoblox.com)</generator><language>en-us</language><lastBuildDate>Mon, 27 Apr 2026 00:00:00 +0000</lastBuildDate><image><url>https://abhijit4-debug.github.io/media/icon_hua5d2116872382d6405c4fb215e02fa8e_2300_512x512_fill_lanczos_center_3.png</url><title>Abhijit Chunduru</title><link>https://abhijit4-debug.github.io/</link></image><item><title>Avoid Forgetting by Preserving Global Knowledge Gradients in Federated Learning with Non-IID Data</title><link>https://abhijit4-debug.github.io/publication/fedproj-tmlr/</link><pubDate>Fri, 01 May 2026 00:00:00 +0000</pubDate><guid>https://abhijit4-debug.github.io/publication/fedproj-tmlr/</guid><description>&lt;p>Work on preserving global knowledge in non-IID federated learning through gradient-projection constraints.&lt;/p></description></item><item><title>Temporally Invertible Latent World Models via Flow Matching</title><link>https://abhijit4-debug.github.io/project/world-models-flow-matching/</link><pubDate>Sun, 01 Mar 2026 00:00:00 +0000</pubDate><guid>https://abhijit4-debug.github.io/project/world-models-flow-matching/</guid><description>&lt;ul>
&lt;li>Introduced temporal cycle consistency in ODE dynamics.&lt;/li>
&lt;li>Improved H16 cosine similarity from 0.247 to 0.87.&lt;/li>
&lt;li>Enabled reverse goal-conditioned planning with 67% success versus 0% baseline.&lt;/li>
&lt;/ul></description></item><item><title>From Thoughts to Trees: Multi-Path Reasoning and Tool Integration in LLMs</title><link>https://abhijit4-debug.github.io/project/llm-reasoning-trees/</link><pubDate>Sat, 01 Nov 2025 00:00:00 +0000</pubDate><guid>https://abhijit4-debug.github.io/project/llm-reasoning-trees/</guid><description>&lt;ul>
&lt;li>Delivered 12-18% accuracy improvement over CoT and ReAct on GSM8K, ARC, and LogiQA.&lt;/li>
&lt;li>Added parallel branch scoring and pruning.&lt;/li>
&lt;li>Reduced incorrect tool calls by 25%.&lt;/li>
&lt;/ul></description></item><item><title>Flotilla: A Scalable, Modular and Resilient Federated Learning Framework for Heterogeneous Resources</title><link>https://abhijit4-debug.github.io/publication/flotilla-jpdc/</link><pubDate>Mon, 01 Sep 2025 00:00:00 +0000</pubDate><guid>https://abhijit4-debug.github.io/publication/flotilla-jpdc/</guid><description>&lt;p>I contributed to the design and development of Flotilla, especially around systems architecture, reliability, and large-scale deployment validation.&lt;/p></description></item><item><title>A Personalized Federated Hypernetworks Based Aggregation Approach for Intrusion Detection Systems</title><link>https://abhijit4-debug.github.io/publication/hypernetworks-ids-scientific-reports/</link><pubDate>Fri, 01 Aug 2025 00:00:00 +0000</pubDate><guid>https://abhijit4-debug.github.io/publication/hypernetworks-ids-scientific-reports/</guid><description>&lt;p>Publication on personalized federated aggregation strategies for intrusion detection systems.&lt;/p></description></item><item><title>Basic Probabilistics of Diffusion Models</title><link>https://abhijit4-debug.github.io/project/diffusion-model-probabilistics/</link><pubDate>Wed, 01 Jan 2025 00:00:00 +0000</pubDate><guid>https://abhijit4-debug.github.io/project/diffusion-model-probabilistics/</guid><description>&lt;ul>
&lt;li>Proved the optimal denoiser as conditional expectation.&lt;/li>
&lt;li>Derived the transport equation via characteristic function analysis.&lt;/li>
&lt;/ul></description></item><item><title>FedCure: A Heterogeneity-Aware Personalized Federated Learning Framework for Intelligent Healthcare in IoMT</title><link>https://abhijit4-debug.github.io/publication/fedcure-ieee-access/</link><pubDate>Mon, 01 Jan 2024 00:00:00 +0000</pubDate><guid>https://abhijit4-debug.github.io/publication/fedcure-ieee-access/</guid><description>&lt;p>Publication on personalized federated learning for intelligent healthcare in IoMT settings.&lt;/p></description></item><item><title>Privacy-Preserving Content Moderation via Federated MapReduce</title><link>https://abhijit4-debug.github.io/project/privacy-preserving-content-moderation/</link><pubDate>Fri, 01 Sep 2023 00:00:00 +0000</pubDate><guid>https://abhijit4-debug.github.io/project/privacy-preserving-content-moderation/</guid><description>&lt;ul>
&lt;li>Reduced training time by 40% across 100+ federated nodes.&lt;/li>
&lt;li>Supported real-time processing while preserving data privacy.&lt;/li>
&lt;/ul></description></item></channel></rss>