-
Safety Aware Task Planning via Large Language Models in Robotics
Authors:
Azal Ahmad Khan,
Michael Andrev,
Muhammad Ali Murtaza,
Sergio Aguilera,
Rui Zhang,
Jie Ding,
Seth Hutchinson,
Ali Anwar
Abstract:
The integration of large language models (LLMs) into robotic task planning has unlocked better reasoning capabilities for complex, long-horizon workflows. However, ensuring safety in LLM-driven plans remains a critical challenge, as these models often prioritize task completion over risk mitigation. This paper introduces SAFER (Safety-Aware Framework for Execution in Robotics), a multi-LLM framewo…
▽ More
The integration of large language models (LLMs) into robotic task planning has unlocked better reasoning capabilities for complex, long-horizon workflows. However, ensuring safety in LLM-driven plans remains a critical challenge, as these models often prioritize task completion over risk mitigation. This paper introduces SAFER (Safety-Aware Framework for Execution in Robotics), a multi-LLM framework designed to embed safety awareness into robotic task planning. SAFER employs a Safety Agent that operates alongside the primary task planner, providing safety feedback. Additionally, we introduce LLM-as-a-Judge, a novel metric leveraging LLMs as evaluators to quantify safety violations within generated task plans. Our framework integrates safety feedback at multiple stages of execution, enabling real-time risk assessment, proactive error correction, and transparent safety evaluation. We also integrate a control framework using Control Barrier Functions (CBFs) to ensure safety guarantees within SAFER's task planning. We evaluated SAFER against state-of-the-art LLM planners on complex long-horizon tasks involving heterogeneous robotic agents, demonstrating its effectiveness in reducing safety violations while maintaining task efficiency. We also verify the task planner and safety planner through actual hardware experiments involving multiple robots and a human.
△ Less
Submitted 19 March, 2025;
originally announced March 2025.
-
Large Language Models in Medical Term Classification and Unexpected Misalignment Between Response and Reasoning
Authors:
Xiaodan Zhang,
Sandeep Vemulapalli,
Nabasmita Talukdar,
Sumyeong Ahn,
Jiankun Wang,
Han Meng,
Sardar Mehtab Bin Murtaza,
Aakash Ajay Dave,
Dmitry Leshchiner,
Dimitri F. Joseph,
Martin Witteveen-Lane,
Dave Chesla,
Jiayu Zhou,
Bin Chen
Abstract:
This study assesses the ability of state-of-the-art large language models (LLMs) including GPT-3.5, GPT-4, Falcon, and LLaMA 2 to identify patients with mild cognitive impairment (MCI) from discharge summaries and examines instances where the models' responses were misaligned with their reasoning. Utilizing the MIMIC-IV v2.2 database, we focused on a cohort aged 65 and older, verifying MCI diagnos…
▽ More
This study assesses the ability of state-of-the-art large language models (LLMs) including GPT-3.5, GPT-4, Falcon, and LLaMA 2 to identify patients with mild cognitive impairment (MCI) from discharge summaries and examines instances where the models' responses were misaligned with their reasoning. Utilizing the MIMIC-IV v2.2 database, we focused on a cohort aged 65 and older, verifying MCI diagnoses against ICD codes and expert evaluations. The data was partitioned into training, validation, and testing sets in a 7:2:1 ratio for model fine-tuning and evaluation, with an additional metastatic cancer dataset from MIMIC III used to further assess reasoning consistency. GPT-4 demonstrated superior interpretative capabilities, particularly in response to complex prompts, yet displayed notable response-reasoning inconsistencies. In contrast, open-source models like Falcon and LLaMA 2 achieved high accuracy but lacked explanatory reasoning, underscoring the necessity for further research to optimize both performance and interpretability. The study emphasizes the significance of prompt engineering and the need for further exploration into the unexpected reasoning-response misalignment observed in GPT-4. The results underscore the promise of incorporating LLMs into healthcare diagnostics, contingent upon methodological advancements to ensure accuracy and clinical coherence of AI-generated outputs, thereby improving the trustworthiness of LLMs for medical decision-making.
△ Less
Submitted 19 December, 2023;
originally announced December 2023.
-
Correct-By-Construction Design of Adaptive Cruise Control with Control Barrier Functions Under Safety and Regulatory Constraints
Authors:
Muhammad Waqas,
Muhammad Ali Murtaza,
Pierluigi Nuzzo,
Petros Ioannou
Abstract:
The safety-critical nature of adaptive cruise control (ACC) systems calls for systematic design procedures, e.g., based on formal methods or control barrier functions (CBFs), to provide strong guarantees of safety and performance under all driving conditions. However, existing approaches have mostly focused on fully verified solutions under smooth traffic conditions, with the exception of stop-and…
▽ More
The safety-critical nature of adaptive cruise control (ACC) systems calls for systematic design procedures, e.g., based on formal methods or control barrier functions (CBFs), to provide strong guarantees of safety and performance under all driving conditions. However, existing approaches have mostly focused on fully verified solutions under smooth traffic conditions, with the exception of stop-and-go scenarios. Systematic methods for high-performance ACC design under safety and regulatory constraints like traffic signals are still elusive. A challenge for correct-by-construction approaches based on CBFs stems from the need to capture the constraints imposed by traffic signals, which lead to candidate time-varying CBFs (TV-CBFs) with finite jump discontinuities in bounded time intervals.
△ Less
Submitted 26 March, 2022;
originally announced March 2022.
-
Distinguishing Number of Non-Zero Component Graphs
Authors:
I. Javaid,
M. Murtaza,
H. Benish
Abstract:
A non-zero component graph $G(\mathbb{V})$ associated to a finite vector space $\mathbb{V}$ is a graph whose vertices are non-zero vectors of $\mathbb{V}$ and two vertices are adjacent, if their corresponding vectors have at least one non-zero component common in their linear combination of basis vectors. In this paper, we extend the study of properties of automorphisms of non-zero component graph…
▽ More
A non-zero component graph $G(\mathbb{V})$ associated to a finite vector space $\mathbb{V}$ is a graph whose vertices are non-zero vectors of $\mathbb{V}$ and two vertices are adjacent, if their corresponding vectors have at least one non-zero component common in their linear combination of basis vectors. In this paper, we extend the study of properties of automorphisms of non-zero component graphs. We prove that every permutation of basis vectors can be extended to an automorphism of $G(\mathbb{V})$. We prove that the symmetric group of basis vectors of $\mathbb{V}$ is isomorphic to the automorphism group of $G(\mathbb{V})$. We find the distinguishing number of the graph for both of the cases, when the number of field elements of vector space $\mathbb{V}$ are 2 or more than 2.
△ Less
Submitted 1 August, 2019;
originally announced August 2019.
-
The Fractional Local Metric Dimension of Graphs
Authors:
Hira Benish,
Muhammad Murtaza,
Imran Javaid
Abstract:
The fractional versions of graph theoretic-invariants multiply the range of applications in scheduling, assignment and operational research problems. In this paper, we introduce the fractional version of local metric dimension of graphs. The local resolving neighborhood $L(xy)$ of an edge $xy$ of a graph $G$ is the set of those vertices in $G$ which resolve the vertices $x$ and $y$. A function…
▽ More
The fractional versions of graph theoretic-invariants multiply the range of applications in scheduling, assignment and operational research problems. In this paper, we introduce the fractional version of local metric dimension of graphs. The local resolving neighborhood $L(xy)$ of an edge $xy$ of a graph $G$ is the set of those vertices in $G$ which resolve the vertices $x$ and $y$. A function $f:V(G)\rightarrow[0, 1]$ is a local resolving function of $G$ if $f(L(xy))\geq1$ for all edges $xy$ in $G$. The minimum value of $f(V(G))$ among all local resolving functions $f$ of $G$ is the fractional local metric dimension of $G$. We study the properties and bounds of fractional local metric dimension of graphs and give some characterization results. We determine the fractional local metric dimension of strong and cartesian product of graphs.
△ Less
Submitted 5 October, 2018;
originally announced October 2018.
-
Automorphism related parameters of a graph associated to a finite vector space
Authors:
Hira Benish,
Imran Javaid,
M. Murtaza
Abstract:
In this paper, we discuss automorphism related parameters of a graph associated to a finite vector space. The fixing neighborhood of a pair $(u,v)$ of vertices of a graph $G$ is the set of all those vertices $w$ of $G$, such that the orbits of $u$ and $v$ under the action of stabilizer of $w$ are not equal. The fixed number of a graph is the minimum number $k$ such that every subset of vertices of…
▽ More
In this paper, we discuss automorphism related parameters of a graph associated to a finite vector space. The fixing neighborhood of a pair $(u,v)$ of vertices of a graph $G$ is the set of all those vertices $w$ of $G$, such that the orbits of $u$ and $v$ under the action of stabilizer of $w$ are not equal. The fixed number of a graph is the minimum number $k$ such that every subset of vertices of $G$ of cardinality $k$ is a fixing set of $G$. We study some properties of automorphisms of a graph associated to finite vector space and find the fixing neighborhood of pair of vertices of the graph. We also find the fixed number of the graph. It is shown that, for every positive integer $N$, there exists a graph $G$ with $fxd(G)-fix(G)\geq N$, where $fxd(G)$ is the fixed number and $fix(G)$ is the fixing number of $G$.
△ Less
Submitted 30 May, 2018; v1 submitted 25 April, 2018;
originally announced April 2018.
-
Locating-Dominating Sets of Functigraphs
Authors:
Muhammad Murtaza,
Muhammad Fazil,
Imran Javaid,
Hira Benish
Abstract:
A locating-dominating set of a graph $G$ is a dominating set of $G$ such that every vertex of $G$ outside the dominating set is uniquely identified by its neighborhood within the dominating set. The location-domination number of $G$ is the minimum cardinality of a locating-dominating set in $G$. Let $G_{1}$ and $G_{2}$ be the disjoint copies of a graph $G$ and $f:V(G_{1})\rightarrow V(G_{2})$ be a…
▽ More
A locating-dominating set of a graph $G$ is a dominating set of $G$ such that every vertex of $G$ outside the dominating set is uniquely identified by its neighborhood within the dominating set. The location-domination number of $G$ is the minimum cardinality of a locating-dominating set in $G$. Let $G_{1}$ and $G_{2}$ be the disjoint copies of a graph $G$ and $f:V(G_{1})\rightarrow V(G_{2})$ be a function. A functigraph $F^f_{G}$ consists of the vertex set $V(G_{1})\cup V(G_{2})$ and the edge set $E(G_{1})\cup E(G_{2})\cup \{uv:v=f(u)\}$. In this paper, we study the variation of the location-domination number in passing from $G$ to $F^f_{G}$ and find its sharp lower and upper bounds. We also study the location-domination number of functigraphs of the complete graphs for all possible definitions of the function $f$. We also obtain the location-domination number of functigraphs of a family of spanning subgraph of the complete graphs.
△ Less
Submitted 15 September, 2017;
originally announced September 2017.
-
Locating-Dominating Sets and Identifying Codes of a Graph Associated to a Finite Vector Space
Authors:
Muhammad Murtaza,
Imran Javaid,
Muhammad Fazil
Abstract:
In this paper, we investigate the problem of covering the vertices of a graph associated to a finite vector space as introduced by Das \cite{Das}, such that we can uniquely identify any vertex by examining the vertices that cover it. We use locating-dominating sets and identifying codes, which are closely related concepts for this purpose. These sets consist of a dominating set of graph such that…
▽ More
In this paper, we investigate the problem of covering the vertices of a graph associated to a finite vector space as introduced by Das \cite{Das}, such that we can uniquely identify any vertex by examining the vertices that cover it. We use locating-dominating sets and identifying codes, which are closely related concepts for this purpose. These sets consist of a dominating set of graph such that every vertex is uniquely identified by its neighborhood within the dominating sets. We find the location-domination number and the identifying number of the graph and study the exchange property for locating-dominating sets and identifying codes.
△ Less
Submitted 30 January, 2017;
originally announced January 2017.
-
On Fixing number of Functigraphs
Authors:
Muhammad Fazil,
Imran Javaid,
Muhammad Murtaza
Abstract:
The fixing number of a graph $G$ is the order of the smallest subset $S$ of its vertex set $V(G)$ such that stabilizer of $S$ in $G$, $Γ_{S}(G)$ is trivial. Let $G_{1}$ and $G_{2}$ be disjoint copies of a graph $G$, and let $g:V(G_{1})\rightarrow V(G_{2})$ be a function. A functigraph $F_{G}$ consists of the vertex set $V(G_{1})\cup V(G_{2})$ and the edge set…
▽ More
The fixing number of a graph $G$ is the order of the smallest subset $S$ of its vertex set $V(G)$ such that stabilizer of $S$ in $G$, $Γ_{S}(G)$ is trivial. Let $G_{1}$ and $G_{2}$ be disjoint copies of a graph $G$, and let $g:V(G_{1})\rightarrow V(G_{2})$ be a function. A functigraph $F_{G}$ consists of the vertex set $V(G_{1})\cup V(G_{2})$ and the edge set $E(G_{1})\cup E(G_{2})\cup \{uv:v=g(u)\}$. In this paper, we study the behavior of the fixing number in passing from $G$ to $F_{G}$ and find its sharp lower and upper bounds. We also study the fixing number of functigraphs of some well known families of graphs like complete graphs, trees and join graphs.
△ Less
Submitted 10 November, 2016;
originally announced November 2016.
-
On The Fixed Number of Graphs
Authors:
I. Javaid,
M. Murtaza,
M. Asif,
F. Iftikhar
Abstract:
An automorphism on a graph $G$ is a bijective mapping on the vertex set $V(G)$, which preserves the relation of adjacency between any two vertices of $G$. An automorphism $g$ fixes a vertex $v$ if $g$ maps $v$ onto itself. The stabilizer of a set $S$ of vertices is the set of all automorphisms that fix vertices of $S$. A set $F$ is called fixing set of $G$, if its stabilizer is trivial. The fixing…
▽ More
An automorphism on a graph $G$ is a bijective mapping on the vertex set $V(G)$, which preserves the relation of adjacency between any two vertices of $G$. An automorphism $g$ fixes a vertex $v$ if $g$ maps $v$ onto itself. The stabilizer of a set $S$ of vertices is the set of all automorphisms that fix vertices of $S$. A set $F$ is called fixing set of $G$, if its stabilizer is trivial. The fixing number of a graph is the cardinality of a smallest fixing set. The fixed number of a graph $G$ is the minimum $k$, such that every $k$-set of vertices of $G$ is a fixing set of $G$. A graph $G$ is called a $k$-fixed graph if its fixing number and fixed number are both $k$. In this paper, we study the fixed number of a graph and give construction of a graph of higher fixed number from graph with lower fixed number. We find bound on $k$ in terms of diameter $d$ of a distance-transitive $k$-fixed graph.
△ Less
Submitted 2 July, 2015;
originally announced July 2015.
-
On some automorphism related parameters in graphs
Authors:
Imran Javaid,
Hira Benish,
Usman Ali,
M. Murtaza
Abstract:
In this paper, we deduce some properties of $f$-sets of connected graphs. Also, we introduce the concept of fixing share of each vertex of a fixing set $D$ to see the participation of each vertex when fixing a connected graph $G$. We define a parameter, called the fixing percentage, by using the concept of fixing share, which is helpful in determining the measure of the amount of fixing done by th…
▽ More
In this paper, we deduce some properties of $f$-sets of connected graphs. Also, we introduce the concept of fixing share of each vertex of a fixing set $D$ to see the participation of each vertex when fixing a connected graph $G$. We define a parameter, called the fixing percentage, by using the concept of fixing share, which is helpful in determining the measure of the amount of fixing done by the elements of $D$ in $G$.
△ Less
Submitted 28 April, 2018; v1 submitted 18 November, 2014;
originally announced November 2014.
-
Weak total resolving sets in graphs
Authors:
Imran Javaid,
Muhammad Salman,
Mahr Murtaza,
Farheen Iftikhar,
Muhammad Imran
Abstract:
A set $W$ of vertices of $G$ is said to be a weak total resolving set for $G$ if $W$ is a resolving set for $G$ as well as for each $w\in W$, there is at least one element in $W-\{w\}$ that resolves $w$ and $v$ for every $v\in V(G)- W$. Weak total metric dimension of $G$ is the smallest order of a weak total resolving set for $G$. This paper includes the investigation of weak total metric dimensio…
▽ More
A set $W$ of vertices of $G$ is said to be a weak total resolving set for $G$ if $W$ is a resolving set for $G$ as well as for each $w\in W$, there is at least one element in $W-\{w\}$ that resolves $w$ and $v$ for every $v\in V(G)- W$. Weak total metric dimension of $G$ is the smallest order of a weak total resolving set for $G$. This paper includes the investigation of weak total metric dimension of trees. Also, weak total resolving number of a graph as well as randomly weak total $k$-dimensional graphs are defined and studied in this paper. Moreover, some characterizations and realizations regarding weak total resolving number and weak total metric dimension are given.
△ Less
Submitted 4 August, 2014;
originally announced August 2014.
