Search

Large language model advancements have enabled the development of multi-agent frameworks to tackle complex, real-world problems such as to automate tasks that require interactions with diverse tools, reasoning, and human collaboration. We present MARCO, a Multi-Agent Real-time Chat Orchestration framework for automating tasks using LLMs. MARCO addresses key challenges in utilizing LLMs for complex, multi-step task execution. It incorporates robust guardrails to steer LLM behavior, validate outputs, and recover from errors that stem from inconsistent output formatting, function and parameter hallucination, and lack of domain knowledge. Through extensive experiments we demonstrate MARCO's superior performance with 94.48% and 92.74% accuracy on task execution for Digital Restaurant Service Platform conversations and Retail conversations datasets respectively along with 44.91% improved latency and 33.71% cost reduction. We also report effects of guardrails in performance gain along with comparisons of various LLM models, both open-source and proprietary. The modular and generic design of MARCO allows it to be adapted for automating tasks across domains and to execute complex usecases through multi-turn interactions., Comment: EMNLP 2024 Industry Track

Knee Osteoarthritis (KOA) is the third most prevalent Musculoskeletal Disorder (MSD) after neck and back pain. To monitor such a severe MSD, a segmentation map of the femur, tibia and tibiofemoral cartilage is usually accessed using the automated segmentation algorithm from the Magnetic Resonance Imaging (MRI) of the knee. But, in recent works, such segmentation is conceivable only from the multistage framework thus creating data handling issues and needing continuous manual inference rendering it unable to make a quick and precise clinical diagnosis. In order to solve these issues, in this paper the Multi-Resolution Attentive-Unet (MtRA-Unet) is proposed to segment the femur, tibia and tibiofemoral cartilage automatically. The proposed work has included a novel Multi-Resolution Feature Fusion (MRFF) and Shape Reconstruction (SR) loss that focuses on multi-contextual information and structural anatomical details of the femur, tibia and tibiofemoral cartilage. Unlike previous approaches, the proposed work is a single-stage and end-to-end framework producing a Dice Similarity Coefficient (DSC) of 98.5% for the femur, 98.4% for the tibia, 89.1% for Femoral Cartilage (FC) and 86.1% for Tibial Cartilage (TC) for critical MRI slices that can be helpful to clinicians for KOA grading. The time to segment MRI volume (160 slices) per subject is 22 sec. which is one of the fastest among state-of-the-art. Moreover, comprehensive experimentation on the segmentation of FC and TC which is of utmost importance for morphology-based studies to check KOA progression reveals that the proposed method has produced an excellent result with binary segmentation

While macroscopic models for single or multi-lane traffic flow are well established, these models are not applicable to the dynamics and characteristics of disordered traffic which is characterized by widely different types of vehicles and no lane discipline. We propose a first-order two-dimensional Lighthill-Whitham-Richards (LWR) model for the continuous macroscopic longitudinal and lateral dynamics of this type of traffic flow. The continuity equation is extended into two dimensions and the equation is closed by assuming a longitudinal flow-density relationship as in traditional one-dimensional models while the lateral dynamics is based on boundary repulsion and a desire of a majority of the drivers to go to less dense regions. This is equivalent to Fick's law giving rise to a lateral diffusion term. Using the proposed model, several numerical tests were conducted under different traffic scenarios representing a wide range of traffic conditions. Even for extreme initial conditions, the model's outcome turned out to be plausible and consistent with observed traffic flow dynamics. Moreover, the numerical convergence test is performed using an analytical solution for lateral steady-state conditions. The model was applied for bicycle simulation and reproduced the evolution of lateral density profile with asymmetric behavior.

Extreme events are unusual and rare large-amplitude fluctuations that occur can unexpectedly in nonlinear dynamical systems. Events above the extreme event threshold of the probability distribution of a nonlinear process characterize extreme events. Different mechanisms for the generation of extreme events and their prediction measures have been reported in the literature. Based on the properties of extreme events, such as rare in frequency of occurrence and extreme in amplitude, various studies have shown that extreme events are both linear and nonlinear in nature. Interestingly, in this work, we report on a special class of extreme events which are nonchaotic and nonperiodic. These nonchaotic extreme events appear in between the quasi-periodic and chaotic dynamics of the system. We report the existence of such extreme events with various statistical measures and characterization techniques., Comment: Accepted for Publication in Physical Review E Letter. 6 Pages, 6 Figures

PostgreSQL is an object-relational database (ORDBMS) that was introduced into the database community and has been avidly used for a variety of information extraction use cases. It is also known to be an advanced SQL-compliant open source Object RDBMS. However, users have not yet resolved to PostgreSQL due to the fact that it is still under the layers and the complexity of its persistent textual environment for an amateur user. Hence, there is a dire need to provide an easy environment for users to comprehend the procedure and standards with which databases are created, tables and the relationships among them, manipulating queries and their flow based on conditions in PostgreSQL. As such, this project identifies the dominant features offered by Postgresql, analyzes the constraints that exist in the database user community in migrating to PostgreSQL and based on the scope and constraints identified, develop a system that will serve as a query generation platform as well as a learning tool that will provide an interactive environment to cognitively learn PostgreSQL query building. This is achieved using a visual editor incorporating a textual editor for a well-versed user. By providing visually-draggable query components to work with, this research aims to offer a cognitive, visual and tactile environment where users can interactively learn PostgreSQL query generation.