\IEEEPARstart{T}{he} past two decades have witnessed an unprecedented accumulation of multi-source remote sensing data from Mars exploration missions. Orbital platforms, such as the Mars Reconnaissance Orbiter, Mars Express, and Tianwen-1, continuously acquire observations across diverse modalities. These modalities range from sub-meter optical imagery (HiRISE) \cite{McEwen24HiRISE} and medium-resolution contextual mosaics (CTX) \cite{Malin07CTX} to hyperspectral mineralogical mapping (CRISM) \cite{Murchie07CRISM} and global topographic models (MOLA) \cite{Smith01MOLA}. Simultaneously, surface assets including the Curiosity \cite{Grotzinger12Curiosity} and Zhurong rovers \cite{Li21ZhuRong} generate complementary in-situ measurements through spectrometers, ground-penetrating radar, and navigation cameras. This rapidly expanding, multi-source, multi-resolution data ecosystem has created a pressing demand for intelligent knowledge retrieval systems that can support planetary scientists in conducting semantic search, cross-source correlation, and multi-scale reasoning over heterogeneous observation archives \cite{Wang26marsretrieval}.
\IEEEPARstart{T}{he} past two decades have witnessed an unprecedented accumulation of multi-source remote sensing data from Mars exploration missions. Orbital platforms, such as the Mars Reconnaissance Orbiter, Mars Express, and Tianwen-1, continuously acquire observations across diverse modalities. These modalities range from sub-meter optical imagery (HiRISE) \cite{McEwen24HiRISE} and medium-resolution contextual mosaics (CTX) \cite{Malin07CTX} to hyperspectral mineralogical mapping (CRISM) \cite{Murchie07CRISM} and global topographic models (MOLA) \cite{Smith01MOLA}. This rapidly expanding, multi-source, multi-resolution data ecosystem has created a pressing demand for intelligent knowledge retrieval systems that can support planetary scientists in conducting semantic search, cross-source correlation, and multi-scale reasoning over heterogeneous observation archives \cite{Wang26marsretrieval}.
Large Language Models (LLMs) have emerged as powerful tools for natural language understanding and generation \cite{Cai25LLM}, and Retrieval Augmented Generation (RAG) has been established as a standard paradigm for grounding LLM responses in external knowledge bases \cite{Lewis20RAG}. By dynamically retrieving relevant documents and conditioning generation on retrieved context, RAG effectively mitigates the hallucination problem inherent in LLMs and enables knowledge-intensive question answering \cite{Zhou24hallucination}. The synergy between LLMs and Knowledge Graphs (KGs) has further advanced retrieval performance through structured knowledge representation, achieving notable improvements in multi-hop reasoning, credibility assessment, and interpretability \cite{Pan24KGandLLM}.
Large Language Models (LLMs) have emerged as powerful tools for natural language understanding and generation \cite{Cai25LLM}, and Retrieval Augmented Generation (RAG) has been established as a standard paradigm for grounding LLM responses in external knowledge bases \cite{Lewis20RAG}. By dynamically retrieving relevant documents and conditioning generation on retrieved context, RAG effectively mitigates the hallucination problem inherent in LLMs and enables knowledge-intensive question answering \cite{Zhou24hallucination}. The synergy between LLMs and Knowledge Graphs (KGs) has further advanced retrieval performance through structured knowledge representation, achieving notable improvements in multi-hop reasoning, credibility assessment, and interpretability \cite{Pan24KGandLLM}.
@@ -60,7 +60,7 @@ Nevertheless, deploying RAG systems for planetary science knowledge retrieval in
\begin{enumerate}
\begin{enumerate}
\item\textbf{The Spatial Topology Loss Problem.} Conventional multi-source retrieval systems judge relevance by textual semantic similarity. Planetary observations are different. Each observation is tied to a spatial footprint on the surface, a time window, and a set of instrument parameters. Two observations are relevant to each other mainly because they are spatially close, temporally overlapping, or captured at complementary resolutions. Existing methods such as multi-source line graphs \cite{Wu25MultiRAG} build graph topology from discrete text entities. This design creates a mismatch with continuous spatial data: $k$ co-located entities need $\binom{k}{2}= O(k^2)$ pairwise edges to represent their spatial relationships. The resulting edge explosion removes the sparsity that these graph models rely on. In short, the discrete graph structure cannot bridge the gap between physical continuity and semantic discreteness.
\item\textbf{The Spatial Topology Loss Problem.} Conventional multi-source retrieval systems judge relevance by textual semantic similarity. Planetary observations are different. Each observation is tied to a spatial footprint on the surface, a time window, and a set of instrument parameters. Two observations are relevant to each other mainly because they are spatially close, temporally overlapping, or captured at complementary resolutions. Existing methods such as multi-source line graphs \cite{Wu25MultiRAG} build graph topology from discrete text entities. This design creates a mismatch with continuous spatial data: $k$ co-located entities need $\binom{k}{2}= O(k^2)$ pairwise edges to represent their spatial relationships. The resulting edge explosion removes the sparsity that these graph models rely on. In short, the discrete graph structure cannot bridge the gap between physical continuity and semantic discreteness.
\item\textbf{The Conflict Over-Smoothing Problem.} Existing multi-source RAG frameworks treat inter-source inconsistency as misinformation or hallucination. They use confidence scores to remove conflicting nodes \cite{Wu25MultiRAG}, \cite{Wang25Astute}. In planetary science, however, different platforms naturally produce different measurements for the same target. An orbiter and a rover observe at different scales, depths, and wavelengths. For example, an orbital spectrometer may detect hydrated minerals on the surface, while an in-situ drill finds olivine-carbonate assemblages below. This conflict does not come from data error. It reflects geological evolution across depth. If we apply uniform conflict filtering, the system suppresses these scientifically valuable signals together with genuine noise. This over-smoothing violates a core principle of deep-space exploration: observational disagreements should be preserved, because they may lead to new discoveries through multi-source comparison.
\item\textbf{The Conflict Over-Smoothing Problem.} Existing multi-source RAG frameworks treat inter-source inconsistency as misinformation or hallucination. They use confidence scores to remove conflicting nodes \cite{Wu25MultiRAG}, \cite{Wang25Astute}. In planetary science, however, different orbital platforms naturally produce different measurements for the same target. Sensors at different wavelengths, spatial resolutions, and viewing angles observe distinct physical aspects of the same surface. For example, a hyperspectral sensor may detect hydrated minerals from spectral absorption features, while a high-resolution optical imager shows no corresponding surface texture change. This conflict does not come from data error. It reflects the multi-dimensional nature of remote sensing observation. If we apply uniform conflict filtering, the system suppresses these scientifically valuable signals together with genuine noise. This over-smoothing violates a core principle of planetary science: observational disagreements between sensors should be preserved, because they may encode genuine physical heterogeneity or lead to new discoveries through multi-source comparison.
\end{enumerate}
\end{enumerate}
To address these two challenges, we propose AreoRAG, a framework designed for multi-source planetary spatial data retrieval augmented generation. AreoRAG introduces two innovations. We first construct a \textbf{Hyperbolic Spatial Hypergraph (HySH)} to resolve the spatial topology loss problem. HySH uses $n$-ary spatial observation hyperedges to group co-located multi-source observations into single high-order facts. This design reduces edge complexity from $O(k^2)$ to $O(k)$. We embed these hyperedges in hyperbolic space via the Lorentz model. The exponential volume growth of negative-curvature geometry naturally fits the hierarchical scale structure of planetary observations. Coarse-resolution global data resides near the origin, while fine-resolution local data extends toward the boundary. To resolve the conflict over-smoothing problem, we develop a \textbf{Physics-Informed Conflict Triage (PICT)} mechanism. PICT replaces uniform conflict filtering with a differentiated triage strategy. It first detects inter-source conflicts through cross-source interaction entropy. Then it classifies each conflict into one of four physically grounded categories: noise, instrument-inherent, scale-dependent, and temporal-evolution. Finally, it applies category-specific confidence recalibration, filtering genuine noise while provably preserving scientifically valuable observational disagreements. The two modules form a tightly coupled loop. HySH provides spatially faithful multi-source evidence to PICT, while PICT feeds back triage results to prioritize scientifically interesting regions in subsequent retrieval.
To address these two challenges, we propose AreoRAG, a framework designed for multi-source planetary spatial data retrieval augmented generation. AreoRAG introduces two innovations. We first construct a \textbf{Hyperbolic Spatial Hypergraph (HySH)} to resolve the spatial topology loss problem. HySH uses $n$-ary spatial observation hyperedges to group co-located multi-source observations into single high-order facts. This design reduces edge complexity from $O(k^2)$ to $O(k)$. We embed these hyperedges in hyperbolic space via the Lorentz model. The exponential volume growth of negative-curvature geometry naturally fits the hierarchical scale structure of planetary observations. Coarse-resolution global data resides near the origin, while fine-resolution local data extends toward the boundary. To resolve the conflict over-smoothing problem, we develop a \textbf{Physics-Informed Conflict Triage (PICT)} mechanism. PICT replaces uniform conflict filtering with a differentiated triage strategy. It first detects inter-source conflicts through cross-source interaction entropy. Then it classifies each conflict into one of four physically grounded categories: noise, instrument-inherent, scale-dependent, and temporal-evolution. Finally, it applies category-specific confidence recalibration, filtering genuine noise while provably preserving scientifically valuable observational disagreements. The two modules form a tightly coupled loop. HySH provides spatially faithful multi-source evidence to PICT, while PICT feeds back triage results to prioritize scientifically interesting regions in subsequent retrieval.
@@ -68,9 +68,9 @@ To address these two challenges, we propose AreoRAG, a framework designed for mu
The contributions of this paper are summarized as follows:
The contributions of this paper are summarized as follows:
\begin{enumerate}
\begin{enumerate}
\item{We propose a Hyperbolic Spatial Hypergraph (HySH) construction module for multi-source planetary data, by combining the $n$-ary hyperedge representation from hypergraph-based RAG \cite{placeholder_HyperRAG} with the Lorentz-model hyperbolic embedding from hyperbolic knowledge graph methods \cite{placeholder_HypRAG}. HySH couples spatial resolution with hyperbolic radial depth so that the hierarchical scale structure of planetary observations is preserved, while edge complexity is reduced from $O(k^2)$ to $O(k)$. We further propose a resolution-aware Spatial Outward Einstein Midpoint (Spatial OEM) aggregation operator with a formal guarantee of outward bias.}
\item{We propose a Hyperbolic Spatial Hypergraph (HySH) construction module for multi-source planetary data, by combining the $n$-ary hyperedge representation from hypergraph-based RAG \cite{lien26hyperrag, luo25hyperrag} with the Lorentz-model hyperbolic embedding from hyperbolic knowledge graph methods \cite{madhu26hyprag}. HySH couples spatial resolution with hyperbolic radial depth so that the hierarchical scale structure of planetary observations is preserved, while edge complexity is reduced from $O(k^2)$ to $O(k)$. We further propose a resolution-aware Spatial Outward Einstein Midpoint (Spatial OEM) aggregation operator with a formal guarantee of outward bias.}
\item{We propose a Physics-Informed Conflict Triage (PICT) mechanism for multi-source retrieval, by adapting the entropy-based conflict detection from \cite{placeholder_TruthfulRAG} and the linear-separability finding of knowledge conflicts from \cite{placeholder_Diagnosing}. PICT classifies each inter-source conflict into four physically grounded categories (noise, instrument-inherent, scale-dependent, temporal-evolution) and applies category-specific confidence recalibration. We provide a formal Anti-Over-Smoothing Guarantee showing that scientifically valuable disagreements are provably preserved. To the best of our knowledge, this is the first conflict-handling mechanism in RAG that explicitly distinguishes erroneous inconsistency from scientifically meaningful observational divergence.}
\item{We propose a Physics-Informed Conflict Triage (PICT) mechanism for multi-source retrieval, by adapting the entropy-based conflict detection from \cite{liu26truthfulrag} and the linear-separability finding of knowledge conflicts from \cite{tang26diagnosing}. PICT classifies each inter-source conflict into four physically grounded categories (noise, instrument-inherent, scale-dependent, temporal-evolution) and applies category-specific confidence recalibration. We provide a formal Anti-Over-Smoothing Guarantee showing that scientifically valuable disagreements are provably preserved. To the best of our knowledge, this is the first conflict-handling mechanism in RAG that explicitly distinguishes erroneous inconsistency from scientifically meaningful observational divergence.}
\item{We design the AreoRAG Prompting (ARP) algorithm that integrates HySH and PICT through three coupling points: spatial alignment as a prerequisite for interaction entropy computation, radial depth difference as a resolution disparity signal for conflict classification, and triage-driven retrieval priority feedback. Experiments on three Mars observation datasets show that AreoRAG outperforms existing multi-source RAG methods in both retrieval accuracy and conflict preservation.}
\item{We design the AreoRAG Prompting (ARP) algorithm that integrates HySH and PICT through three coupling points: spatial alignment as a prerequisite for interaction entropy computation, radial depth difference as a resolution disparity signal for conflict classification, and triage-driven retrieval priority feedback. Experiments on three Mars observation datasets show that AreoRAG outperforms existing multi-source RAG methods in both retrieval accuracy and conflict preservation.}
\end{enumerate}
\end{enumerate}
@@ -94,7 +94,8 @@ where $\text{LLM}(q_i, d_l)$ denotes the relevance score between query $q_i$ and
Furthermore, we optimize the knowledge construction and retrieval modules by introducing a hyperbolic spatial hypergraph to achieve spatially faithful knowledge aggregation and physics-informed conflict handling. Specifically, the proposed approach is formally defined through the following definitions.
Furthermore, we optimize the knowledge construction and retrieval modules by introducing a hyperbolic spatial hypergraph to achieve spatially faithful knowledge aggregation and physics-informed conflict handling. Specifically, the proposed approach is formally defined through the following definitions.
Definition~1 (Multi-source planetary observation data). Given a set of observation platforms $\mathcal{H}$ (e.g., MRO, Mars Express, Tianwen-1, Curiosity, Zhurong), the observation data $D =\{\mathcal{I}, \mathcal{P}_{foot}, \mathcal{T}_{win}, \mathcal{S}_{band}, c, \text{meta}\}$ exists, where $\mathcal{I}$ denotes the instrument identity, $\mathcal{P}_{foot}\subset\mathbb{S}^2_{Mars}$ denotes the spatial footprint on the Martian surface, $\mathcal{T}_{win}$ denotes the temporal acquisition window parameterized by Solar Longitude $L_s$, $\mathcal{S}_{band}$ denotes the spectral band configuration, $c$ represents the observation content (image, spectrum, or derived product), and meta represents the PDS/CNSA metadata. Through a multi-source spatial adapter parsing algorithm, we obtain normalized data $\widehat{D}=\{\text{id}, \mathcal{I}, \mathcal{P}_{foot}, \mathcal{T}_{win}, \mathcal{S}_{band}, \ell_{res}, \text{jsc}, \text{meta}\}$, where id is the unique identifier, $\ell_{res}\in\mathbb{R}^+$ denotes the ground sampling distance (spatial resolution), and jsc denotes the observation content stored using JSON-LD for linked data interoperability.
% TODO：改成遥感数据
Definition~1 (Multi-source planetary observation data). Given a set of orbital observation platforms $\mathcal{H}$ (e.g., MRO, Mars Express, Tianwen-1), the observation data $D =\{\mathcal{I}, \mathcal{P}_{foot}, \mathcal{T}_{win}, \mathcal{S}_{band}, c, \text{meta}\}$ exists, where $\mathcal{I}$ denotes the instrument identity, $\mathcal{P}_{foot}\subset\mathbb{S}^2_{Mars}$ denotes the spatial footprint on the Martian surface, $\mathcal{T}_{win}$ denotes the temporal acquisition window parameterized by Solar Longitude $L_s$, $\mathcal{S}_{band}$ denotes the spectral band configuration, $c$ represents the observation content (image, spectrum, or derived product), and meta represents the PDS/CNSA metadata. Through a multi-source spatial adapter parsing algorithm, we obtain normalized data $\widehat{D}=\{\text{id}, \mathcal{I}, \mathcal{P}_{foot}, \mathcal{T}_{win}, \mathcal{S}_{band}, \ell_{res}, \text{jsc}, \text{meta}\}$, where id is the unique identifier, $\ell_{res}\in\mathbb{R}^+$ denotes the ground sampling distance (spatial resolution), and jsc denotes the observation content stored using JSON-LD for linked data interoperability.
Definition~2 ($N$-ary spatial knowledge hypergraph). An $n$-ary spatial knowledge hypergraph is defined as $\mathcal{G}_{hyp}=(\mathcal{E}, \mathcal{R}, \mathcal{F}_{spa})$, where $\mathcal{E}$ denotes the entity set, $\mathcal{R}$ denotes the relation set, and $\mathcal{F}_{spa}$ denotes the set of spatial observation hyperedges. Each spatial observation hyperedge $f_{spa}^n \in\mathcal{F}_{spa}$ binds multiple entities and observation parameters into a single $n$-ary relational fact:
Definition~2 ($N$-ary spatial knowledge hypergraph). An $n$-ary spatial knowledge hypergraph is defined as $\mathcal{G}_{hyp}=(\mathcal{E}, \mathcal{R}, \mathcal{F}_{spa})$, where $\mathcal{E}$ denotes the entity set, $\mathcal{R}$ denotes the relation set, and $\mathcal{F}_{spa}$ denotes the set of spatial observation hyperedges. Each spatial observation hyperedge $f_{spa}^n \in\mathcal{F}_{spa}$ binds multiple entities and observation parameters into a single $n$-ary relational fact:
@@ -116,21 +117,22 @@ Definition~4 (Observation-grounded homologous data). For a query $Q(q, \mathcal{
Definition~5 (Observation-grounded knowledge source). A planetary observation knowledge source is defined as $\mathcal{K}_s =(\mathcal{I}_s, \Omega_s, F(\mathcal{K}_s), \mathcal{M}_s)$, where $\mathcal{I}_s$ denotes the instrument, $\Omega_s =(\ell_{res}, \lambda_{band}, \theta_{view}, d_{pen})$ denotes the observation geometry parameters (spatial resolution, spectral band, viewing angle, penetration depth), $F(\mathcal{K}_s)$ denotes the set of atomic factual statements, and $\mathcal{M}_s$ denotes the physical measurement model that maps target properties through observation constraints to observable facts.
Definition~5 (Observation-grounded knowledge source). A planetary observation knowledge source is defined as $\mathcal{K}_s =(\mathcal{I}_s, \Omega_s, F(\mathcal{K}_s), \mathcal{M}_s)$, where $\mathcal{I}_s$ denotes the instrument, $\Omega_s =(\ell_{res}, \lambda_{band}, \theta_{view}, d_{pen})$ denotes the observation geometry parameters (spatial resolution, spectral band, viewing angle, penetration depth), $F(\mathcal{K}_s)$ denotes the set of atomic factual statements, and $\mathcal{M}_s$ denotes the physical measurement model that maps target properties through observation constraints to observable facts.
Definition 6 (Conflict triage confidence.) For observation-grounded homologous data obtained from the spatial hypergraph, the conflict triage confidence integrates two levels of assessment: (a) cross-source interaction entropy to detect inter-source conflicts, and (b) physics-informed conflict classification to determine whether detected conflicts represent noise to be filtered or scientifically meaningful observational divergences to be preserved. Unlike conventional candidate confidence [14] that uniformly penalizes inconsistency, conflict triage confidence applies differentiated recalibration based on the physical origin of each conflict.
% TODO Unlike conventional candidate confidence \cite{Wu25MultiRAG}，传统的用词不准确
Definition 6 (Conflict triage confidence). For observation-grounded homologous data obtained from the spatial hypergraph, the conflict triage confidence integrates two levels of assessment: (a) cross-source interaction entropy to detect inter-source conflicts, and (b) physics-informed conflict classification to determine whether detected conflicts represent noise to be filtered or scientifically meaningful observational divergences to be preserved. Unlike conventional candidate confidence \cite{Wu25MultiRAG} that uniformly penalizes inconsistency, conflict triage confidence applies differentiated recalibration based on the physical origin of each conflict.
\section{Methodology}
\section{Methodology}
% TODO 要有一张总图
\subsection{Framework of AreoRAG}
\subsection{Framework of AreoRAG}
This section introduces the implementation approach of AreoRAG. As shown in Fig. 3, the framework comprises three coupled modules. The first step involves constructing a hyperbolic spatial hypergraph (HySH) from multi-source planetary observation data, achieving unified spatiotemporal representation via $n$-ary observation hyperedges embedded in hyperbolic space (Sec.~\ref{sec:HySH}); the second step performs spatiotemporal retrieval on the constructed HySH, where hyperbolic spatial proximity encoding and cross-resolution aggregation via the Spatial Outward Einstein Midpoint are employed to extract query-relevant multi-source evidence (Sec.~\ref{sec:retrieval}); the third step applies Physics-Informed Conflict Triage (PICT), which detects inter-source conflicts via cross-source interaction entropy, classifies them into four scientific categories, and executes conflict-aware confidence recalibration to preserve scientifically valuable disagreements while filtering noise (Sec.~\ref{sec:PICT}). Finally, integrating the aforementioned steps to form the AreoRAG Prompting algorithm (ARP) (Sec.~\ref{sec:prompt}).
This section elaborates on the implementation approach of AreoRAG. As shown in Fig. 3, the framework comprises three tightly coupled modules. The first step involves constructing a Hyperbolic Spatial Hypergraph (HySH) from multi-source planetary observation data, achieving unified spatiotemporal representation via $n$-ary observation hyperedges embedded in hyperbolic space (Section III-B); the second step performs spatiotemporal retrieval on the constructed HySH, where hyperbolic spatial proximity encoding and cross-resolution aggregation via the Spatial Outward Einstein Midpoint are employed to extract query-relevant multi-source evidence (Section III-C); the third step applies Physics-Informed Conflict Triage (PICT), which detects inter-source conflicts via cross-source interaction entropy, classifies them into four scientific categories, and executes conflict-aware confidence recalibration to preserve scientifically valuable disagreements while filtering noise (Section III-D). Finally, integrating the aforementioned steps to form the AreoRAG Prompting algorithm, ARP (Section III-E).
The three modules interact through three explicit coupling points: (1) HySH's spatial alignment is a prerequisite for meaningful interaction entropy computation in PICT; (2) the radial depth difference $\Delta r$ from HySH directly feeds into the PICT feature vector as the resolution disparity signal; and (3) PICT's triage results feed back to boost retrieval priority of scientifically interesting regions in subsequent queries.
The three modules interact through three explicit coupling points: (1) HySH's spatial alignment is a prerequisite for meaningful interaction entropy computation in PICT; (2) the radial depth difference $\Delta r$ from HySH directly feeds into the PICT feature vector as the resolution disparity signal; and (3) PICT's triage results feed back to boost retrieval priority of scientifically interesting regions in subsequent queries.
The AreoRAG method begins by constructing a knowledge structure that can faithfully represent the continuous spatiotemporal topology of planetary multi-source data. Unlike MultiRAG's Multi-source Line Graph (MLG), which relies on discrete text entities and binary triples, we introduce a hypergraph structure embedded in hyperbolic space to jointly address edge explosion and spatial scale hierarchy.
The AreoRAG method begins by constructing a knowledge structure that can faithfully represent the continuous spatiotemporal topology of planetary multi-source data. Unlike MultiRAG's Multi-source Line Graph (MLG), which relies on discrete text entities and binary triples, we introduce a hypergraph structure embedded in hyperbolic space to jointly address edge explosion and spatial scale hierarchy.
1) Multi-source Spatial Adapter Parsing: We first design a spatial adapter for each observation data source to parse instrument metadata, spatial footprints, temporal windows, and spectral parameters. For orbital remote sensing data (e.g., HiRISE, CTX, CRISM), parsing involves extracting the image footprint geometry, ground sampling distance, and spectral band configuration from PDS labels. For in-situ data (e.g., rover spectrometers, ground-penetrating radar), parsing extracts the rover traverse coordinates, measurement timestamps in Sol, and instrument-specific parameters such as penetration depth. All temporal references are unified to Solar Longitude $L_s$ to enable cross-platform temporal comparison. For derived data products (e.g., DTMs, mineral abundance maps), parsing extracts provenance links to the source observations and processing parameters.
1) Multi-source Spatial Adapter Parsing: We first design a spatial adapter for each observation data source to parse instrument metadata, spatial footprints, temporal windows, and spectral parameters. For orbital remote sensing data (e.g., HiRISE, CTX, CRISM, MOLA), parsing involves extracting the image footprint geometry, ground sampling distance, and spectral band configuration from PDS labels. For derived data products (e.g., DTMs, mineral abundance maps), parsing extracts provenance links to the source observations and processing parameters. All temporal references are unified to Solar Longitude $L_s$ to enable cross-platform temporal comparison.
The final integration of multi-source spatial data can be expressed as:
The final integration of multi-source spatial data can be expressed as:
@@ -138,7 +140,7 @@ The final integration of multi-source spatial data can be expressed as:
\label{equ:multi-source spatial data}
\label{equ:multi-source spatial data}
D_{Fusion} = \bigcup_{i=1}^{n} A_i^{spa}(D_i),
D_{Fusion} = \bigcup_{i=1}^{n} A_i^{spa}(D_i),
\end{equation}
\end{equation}
where $A_i^{spa}\in\{Ada_{orbital}, Ada_{insitu}, Ada_{derived}\}$ represents the spatial adapter parsing functions for orbital, in-situ, and derived data products respectively, and $D_i$ represents the original observation datasets from different platforms.
where $A_i^{spa}\in\{Ada_{orbital}, Ada_{derived}\}$ represents the spatial adapter parsing functions for orbital and derived data products respectively, and $D_i$ represents the original observation datasets from different platforms.
Through the parsed data $D_{Fusion}$, we further extract entities (geological features, mineral signatures, topographic structures), relationships (spatial containment, temporal succession, compositional association), and observation-specific attributes. The knowledge extraction process employs LLM-based entity recognition guided by a planetary science domain schema:
Through the parsed data $D_{Fusion}$, we further extract entities (geological features, mineral signatures, topographic structures), relationships (spatial containment, temporal succession, compositional association), and observation-specific attributes. The knowledge extraction process employs LLM-based entity recognition guided by a planetary science domain schema:
Through this embedding, global coarse-resolution data (e.g., MOLA topography at ~460 m) is placed near the hyperbolic origin (small radial depth), while local high-resolution data (e.g., HiRISE at 0.3 m) is placed far from the origin (large radial depth). The exponential volume growth of $\mathbb{H}_K^d$ naturally accommodates the exponentially increasing number of observations at finer scales.
Through this embedding, global coarse-resolution data (e.g., MOLA topography at ~460 m) is placed near the hyperbolic origin (small radial depth), while local high-resolution data (e.g., HiRISE at 0.3 m) is placed far from the origin (large radial depth). The exponential volume growth of $\mathbb{H}_K^d$ naturally accommodates the exponentially increasing number of observations at finer scales.
4) Cross-Reference-Frame Alignment: To address the heterogeneous reference frame problem (orbiter areocentric coordinates vs. rover-centric local coordinates), we align all observations to a global reference via parallel transport on the hyperbolic manifold:
4) Cross-Reference-Frame Alignment: Different orbital missions use slightly different coordinate reference frames. We align all observations to a unified global reference via parallel transport on the hyperbolic manifold:
\begin{equation}
\begin{equation}
\label{equ:Cross-Reference-Frame Alignment}
\label{equ:Cross-Reference-Frame Alignment}
@@ -178,10 +180,10 @@ Through this embedding, global coarse-resolution data (e.g., MOLA topography at
\end{equation}
\end{equation}
where $\log_{o_k}$ is the logarithmic map at the local reference origin $o_k$, $\Gamma_{o_k \to o_g}$ is the parallel transport operator along the geodesic from $o_k$ to the global origin $o_g$, and $\exp_{o_g}$ is the exponential map at the global origin. Unlike Euclidean affine transformations, hyperbolic parallel transport preserves geodesic distances and radial depth, ensuring that scale hierarchy information is maintained after cross-frame alignment.
where $\log_{o_k}$ is the logarithmic map at the local reference origin $o_k$, $\Gamma_{o_k \to o_g}$ is the parallel transport operator along the geodesic from $o_k$ to the global origin $o_g$, and $\exp_{o_g}$ is the exponential map at the global origin. Unlike Euclidean affine transformations, hyperbolic parallel transport preserves geodesic distances and radial depth, ensuring that scale hierarchy information is maintained after cross-frame alignment.
Here, we provide a simple example of hyperbolic spatial hypergraph construction. As shown in Fig. 4, an observation region is covered by three sources at different resolutions: a CTX mosaic (6 m), an HiRISE strip (0.3 m), and a CRISM spectral cube (18 m). In the HySH, the HiRISE observation (finest resolution) is embedded at the largest radial depth, while the CRISM observation (coarsest resolution) is nearest to the origin. A spatial observation hyperedge binds all three observations and their co-located geological features into a single $n$-ary fact, without requiring $O(k^2)$ pairwise edges.
Here, we provide a simple example of hyperbolic spatial hypergraph construction. As shown in Fig. 4, an observation region is covered by three orbital sensors at different resolutions: a CTX mosaic (6 m), an HiRISE strip (0.3 m), and a CRISM spectral cube (18 m). In the HySH, the HiRISE observation (finest resolution) is embedded at the largest radial depth, while the CRISM observation (coarsest resolution) is nearest to the origin. A spatial observation hyperedge binds all three observations and their co-located geological features into a single $n$-ary fact, without requiring $O(k^2)$ pairwise edges.
\subsection{Spatiotemporal Retrieval with Cross-Resolution Aggregation}
\subsection{Spatiotemporal Retrieval with Cross-Resolution Aggregation}\label{sec:retrieval}
After the construction of the hyperbolic spatial hypergraph, the next step is to retrieve query-relevant multi-source spatial evidence. The retrieval process comprises two phases: spatiotemporal evidence extraction and cross-resolution aggregation.
After the construction of the hyperbolic spatial hypergraph, the next step is to retrieve query-relevant multi-source spatial evidence. The retrieval process comprises two phases: spatiotemporal evidence extraction and cross-resolution aggregation.
The outward bias guarantees that high-resolution observations dominate the aggregated representation. This is essential for planetary science retrieval: when a user queries a specific geological feature, the aggregated evidence should preserve the fine-scale observational details rather than being smoothed into a coarse-resolution summary.
The outward bias guarantees that high-resolution observations dominate the aggregated representation. This is essential for planetary science retrieval: when a user queries a specific geological feature, the aggregated evidence should preserve the fine-scale observational details rather than being smoothed into a coarse-resolution summary.
We define the multi-source spatial evidence retrieved in a single query as observation-grounded homologous data (Definition 4). Although targeting the same query object, these data often provide inconsistent factual statements due to differences in instrument principles, observation geometry, and acquisition epochs. Unlike MultiRAG's Multi-level Confidence Computing (MCC), which assumes that inconsistency indicates unreliability and employs mutual information entropy to filter conflicting nodes, we adopt a fundamentally different paradigm: Physics-Informed Conflict Triage (PICT), which classifies conflicts by their physical origin and applies differentiated processing strategies.
We define the multi-source spatial evidence retrieved in a single query as observation-grounded homologous data (Definition 4). Although targeting the same query object, these data often provide inconsistent factual statements due to differences in instrument principles, observation geometry, and acquisition epochs. Unlike MultiRAG's Multi-level Confidence Computing (MCC), which assumes that inconsistency indicates unreliability and employs mutual information entropy to filter conflicting nodes, we adopt a fundamentally different paradigm: Physics-Informed Conflict Triage (PICT), which classifies conflicts by their physical origin and applies differentiated processing strategies.
This theorem provides a formal guarantee that scientifically valuable conflict nodes can never be suppressed below their baseline confidence by the triage mechanism, directly addressing the over-smoothing problem.
This theorem provides a formal guarantee that scientifically valuable conflict nodes can never be suppressed below their baseline confidence by the triage mechanism, directly addressing the over-smoothing problem.
\subsection{AreoRAG Prompting}
\subsection{AreoRAG Prompting}\label{sec:prompt}
We propose the AreoRAG Prompting (ARP) algorithm for multi-source planetary spatial data retrieval. The complete procedure is presented in Algorithm~\ref{alg:arp}.
We propose the AreoRAG Prompting (ARP) algorithm for multi-source planetary spatial data retrieval. The complete procedure is presented in Algorithm~\ref{alg:arp}.
@@ -357,9 +359,9 @@ This section conducts experiments and performance analysis on the Hyperbolic Spa
**a) Datasets:** To validate the effectiveness of AreoRAG in planetary multi-source spatial data retrieval, we construct three datasets from real Mars exploration archives and further evaluate on two general multi-hop QA benchmarks. The planetary datasets are summarized in Table I.
**a) Datasets:** To validate the effectiveness of AreoRAG in planetary multi-source spatial data retrieval, we construct three datasets from real Mars exploration archives and further evaluate on two general multi-hop QA benchmarks. The planetary datasets are summarized in Table I.
(1) **MarsRegion-QA**: A multi-source spatial QA dataset constructed from the Mars Orbital Data Explorer (ODE) archives. We select five scientifically significant regions on Mars — Jezero Crater, Gale Crater, Utopia Planitia (Zhurong landing site), Valles Marineris, and Olympus Mons — and aggregate observations from HiRISE (0.3 m), CTX (6 m), CRISM (18 m), MOLA (460 m), and Zhurong/Curiosity rover in-situ measurements. Each query targets cross-source spatial reasoning (e.g., "What mineral signatures have been detected in the clay-bearing unit at the western delta of Jezero Crater, and do orbital and in-situ observations agree?"). We construct 200 queries with expert-annotated ground truth answers and conflict labels.
(1) **MarsRegion-QA**: A multi-source spatial QA dataset constructed from the Mars Orbital Data Explorer (ODE) archives. We select five scientifically significant regions on Mars — Jezero Crater, Gale Crater, Utopia Planitia, Valles Marineris, and Olympus Mons — and aggregate orbital observations from HiRISE (0.3 m), CTX (6 m), CRISM (18 m), and MOLA (460 m). Each query targets cross-source spatial reasoning (e.g., "What mineral signatures have been detected in the clay-bearing unit at the western delta of Jezero Crater, and do different orbital sensors agree?"). We construct 200 queries with expert-annotated ground truth answers and conflict labels.
(2) **MarsConflict-50**: A curated subset of 50 observation pairs exhibiting known scientific conflicts documented in the planetary science literature (e.g., orbital detection of hydrated minerals vs. inconclusive in-situ results). Each pair is annotated with conflict type (instrument-inherent, scale-dependent, temporal-evolution, or noise) by domain experts. This dataset serves as the primary benchmark for evaluating PICT's conflict classification accuracy.
(2) **MarsConflict-50**: A curated subset of 50 observation pairs exhibiting known scientific conflicts documented in the planetary science literature (e.g., CRISM detection of hydrated minerals vs. contradictory results from other spectral sensors at the same location). Each pair is annotated with conflict type (instrument-inherent, scale-dependent, temporal-evolution, or noise) by domain experts. This dataset serves as the primary benchmark for evaluating PICT's conflict classification accuracy.
(3) **MarsTemporal-QA**: A temporal reasoning dataset comprising 150 queries about surface changes observed across different Mars Years (MY), such as recurring slope lineae (RSL) activity, dust storm impacts, and seasonal frost patterns. Each query requires integrating observations spanning $L_s$ ranges to assess temporal evolution.
(3) **MarsTemporal-QA**: A temporal reasoning dataset comprising 150 queries about surface changes observed across different Mars Years (MY), such as recurring slope lineae (RSL) activity, dust storm impacts, and seasonal frost patterns. Each query requires integrating observations spanning $L_s$ ranges to assess temporal evolution.
@@ -374,19 +376,17 @@ This section conducts experiments and performance analysis on the Hyperbolic Spa
@@ -629,7 +629,7 @@ AreoRAG's effectiveness in multi-source planetary data integration is demonstrat
This case study exemplifies AreoRAG's core advantage: while MultiRAG filters the in-situ observation as "unreliable" due to its inconsistency with orbital data, AreoRAG recognizes this as a scale-dependent conflict, preserves both observations, and generates a scientifically meaningful explanation (spatial mixing effect). The answer includes provenance metadata (DataIDs) for scientific traceability, and proactively recommends follow-up data to resolve the ambiguity — a capability enabled by the PICT module's conflict-aware context construction.
This case study exemplifies AreoRAG's core advantage: while MultiRAG filters one of the conflicting orbital observations as "unreliable," AreoRAG recognizes this as a scale-dependent conflict between sensors operating at different resolutions, preserves both observations, and generates a scientifically meaningful explanation (spatial mixing effect at different scales). The answer includes provenance metadata (DataIDs) for scientific traceability, and proactively recommends follow-up data to resolve the ambiguity — a capability enabled by the PICT module's conflict-aware context construction.
\subsection{Limitations}
\subsection{Limitations}
@@ -687,7 +687,7 @@ In the broader geospatial domain, the integration of AI with remote sensing data
To the best of our knowledge, AreoRAG is the first framework that brings RAG capabilities to planetary remote sensing data retrieval. By constructing a spatially-grounded knowledge hypergraph with physics-informed conflict handling, AreoRAG transforms the planetary data retrieval paradigm from metadata keyword matching to semantic spatial reasoning, enabling natural language queries that involve spatial proximity, temporal evolution, cross-source correlation, and scientifically informed conflict interpretation.
To the best of our knowledge, AreoRAG is the first framework that brings RAG capabilities to planetary remote sensing data retrieval. By constructing a spatially-grounded knowledge hypergraph with physics-informed conflict handling, AreoRAG transforms the planetary data retrieval paradigm from metadata keyword matching to semantic spatial reasoning, enabling natural language queries that involve spatial proximity, temporal evolution, cross-source correlation, and scientifically informed conflict interpretation.
\section{Conclusion}
\section{Conclusion}
In this work, we introduce AreoRAG, a framework designed for multi-source planetary spatial data retrieval augmented generation. To address the structural bottleneck of discrete representation failure for continuous spatiotemporal topology and the epistemological conflict between scientific observational divergence and traditional de-falsification mechanisms, we propose two key innovations: Hyperbolic Spatial Hypergraph construction and Physics-Informed Conflict Triage.
In this work, we introduce AreoRAG, a framework designed for multi-source planetary remote sensing data retrieval augmented generation. To address the spatial topology loss problem of discrete graph representations and the conflict over-smoothing problem of existing de-falsification mechanisms, we propose two key innovations: Hyperbolic Spatial Hypergraph construction and Physics-Informed Conflict Triage.
The introduction of HySH employs $n$-ary spatial observation hyperedges embedded in hyperbolic space via the Lorentz model, reducing edge complexity from $O(k^2)$ to $O(k)$ while faithfully preserving the hierarchical scale structure of planetary observations through the scale-curvature correspondence principle. The Spatial Outward Einstein Midpoint aggregation operator further ensures that cross-resolution evidence fusion retains fine-scale observational details with a formal outward bias guarantee. Meanwhile, the PICT module fundamentally redefines the role of inter-source conflict in RAG systems — shifting from uniform conflict elimination to physics-informed conflict triage that classifies disagreements by their physical origin and applies differentiated confidence recalibration. The Anti-Over-Smoothing Guarantee (Theorem 2) ensures that scientifically valuable observational divergences are provably preserved rather than suppressed.
The introduction of HySH employs $n$-ary spatial observation hyperedges embedded in hyperbolic space via the Lorentz model, reducing edge complexity from $O(k^2)$ to $O(k)$ while faithfully preserving the hierarchical scale structure of planetary observations through the scale-curvature correspondence principle. The Spatial Outward Einstein Midpoint aggregation operator further ensures that cross-resolution evidence fusion retains fine-scale observational details with a formal outward bias guarantee. Meanwhile, the PICT module fundamentally redefines the role of inter-source conflict in RAG systems — shifting from uniform conflict elimination to physics-informed conflict triage that classifies disagreements by their physical origin and applies differentiated confidence recalibration. The Anti-Over-Smoothing Guarantee (Theorem 2) ensures that scientifically valuable observational divergences are provably preserved rather than suppressed.
title={HyperRAG: Reasoning N-ary Facts over Hypergraphs for Retrieval Augmented Generation},
author={Lien, Wen-Sheng and Chan, Yu-Kai and Hsiao, Hao-Lung and Ruan, Bo-Kai and Chiang, Meng-Fen and Chen, Chien-An and Yeh, Yi-Ren and Shuai, Hong-Han},
journal={arXiv preprint arXiv:2602.14470},
year={2026}
}
@article{luo25hyperrag,
title={Hypergraphrag: Retrieval-augmented generation via hypergraph-structured knowledge representation},
author={Luo, Haoran and Chen, Guanting and Zheng, Yandan and Wu, Xiaobao and Guo, Yikai and Lin, Qika and Feng, Yu and Kuang, Zemin and Song, Meina and Zhu, Yifan and others},
journal={arXiv preprint arXiv:2503.21322},
year={2025}
}
@article{madhu26hyprag,
title={HypRAG: Hyperbolic Dense Retrieval for Retrieval Augmented Generation},
author={Madhu, Hiren and Bui, Ngoc and Maatouk, Ali and Tassiulas, Leandros and Krishnaswamy, Smita and Yang, Menglin and Ganguly, Sukanta and Srinivasan, Kiran and Ying, Rex},
journal={arXiv preprint arXiv:2602.07739},
year={2026}
}
@inproceedings{liu26truthfulrag,
title={Truthfulrag: Resolving factual-level conflicts in retrieval-augmented generation with knowledge graphs},
author={Liu, Shuyi and Shang, Yu-Ming and Zhang, Xi},
booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
volume={40},
number={38},
pages={32168--32176},
year={2026}
}
@article{tang26diagnosing,
title={Diagnosing Knowledge Conflict in Multimodal Long-Chain Reasoning},
author={Tang, Jing and Wang, Kun and Lu, Haolang and Chen, Hongjin and Chen, KaiTao and Sun, Zhongxiang and Li, Qiankun and Lyu, Lingjuan and Nan, Guoshun and Zeng, Zhigang},
journal={arXiv preprint arXiv:2602.14518},
year={2026}
}
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
龙澳