用稀疏自编码器拆解 AlphaEarth 地球观测嵌入，让卫星特征可解释
Decomposing Earth Embeddings with Sparse Autoencoders
给做遥感或可解释性研究的人看的：作者把 AlphaEarth、Tessera 这类地球观测嵌入拆成了可读的稀疏特征，还配了一套专门评估协议。发现对齐概念数从维度中位数 12 个降到接近 1 个，细尺度方差占比提升到 23%–32%。想给卫星模型做可解释性分析的话，这篇给出了方法起点。
arXiv 论文提出用稀疏自编码器（SAE）将地球观测基础模型的嵌入分解为稀疏、近似单语义的特征，研究对象包括 AlphaEarth 和 Tessera 等全球米级分辨率嵌入模型。
由于 LLM 判断对卫星像素不适用且空间自相关显著，作者设计了专门的评估协议，从匹配稀疏度下的重建精度、与参考概念的对齐程度、空间尺度三个维度打分。结果显示 SAE 特征语义更纯净：AlphaEarth 上 88% 的特征最多对齐 20 个概念组中的 1 个，而原始嵌入维度中位数对齐 12 个；BatchTopK 和 Matryoshka 变体在 640 米 patch 内承载 23%–32% 的细尺度方差，远高于原始维度的 9%。
实验还发现冻结随机字典的 SAE 同样具备这两项性质，说明其源于稀疏性本身，但只有学习到的字典能准确重建嵌入——AlphaEarth 上训练后的 SAE 未解释方差比基线少 8 倍。论文测试了土地覆盖子概念发现和分布外像素检测两个应用，目前可靠性尚不足以实用。
Decomposing Earth Embeddings with Sparse Autoencoders
Earth observation foundation models produce global metre-resolution embeddings that underpin work across agriculture, ecology, hydrology, and beyond. Yet these embeddings are opaque: each dimension entangles unrelated physical concepts, which makes them difficult to interpret and their robustness hard to assess. Here, we propose sparse autoencoders (SAEs) as a principled way to decompose Earth embeddings into sparse, approximately monosemantic features over an overcomplete dictionary. Because LLM judges of interpretations do not transfer to satellite pixels, and spatial autocorrelation matters, we introduce a tailored evaluation protocol. It scores any SAE on reconstruction at matched sparsity, alignment with reference concepts, and spatial scale, each against baselines that include an SAE whose decoder is a frozen random dictionary. On this protocol, SAEs reconstruct AlphaEarth and Tessera embeddings better than every baseline at matched sparsity. Their features align with far fewer reference concepts than raw embedding dimensions (on AlphaEarth, 88% align with at most one of 20 concept groups, most with none, against a median of 12 per raw dimension). They also carry more of their variance at fine spatial scales (23 to 32% within a 640 m patch for BatchTopK and Matryoshka, against 9%). An SAE with a frozen random decoder shares both properties, which therefore follow from sparsity, whereas only a learned dictionary reconstructs the embedding accurately: on AlphaEarth, trained SAEs leave eight times less variance unexplained. We test the sparse code on two applications, discovering land-cover sub-concepts and detecting out-of-distribution pixels by area-of-applicability estimation. Both indicate how sparse features could be used, but neither is yet reliable enough for practical use. Together, these results are a step toward Earth observation models whose features can be inspected.
本条情报归属 AITOP「论文」核心追踪栏目。在跨区域产业观察坐标中，该动态直接关联「AI 开发者工具链、Coding Agent 与工程化实践」研究主线，反映出当前技术演进与商业落地的最新风向。涉及核心实体与技术议题包括：AlphaEarth、Tessera、稀疏自编码器、可解释性，读者可通过下方标签进入对应专题页查阅历史脉络与关联对比。
信源核验说明：本文原始线索采编自「arXiv cs.LG」（@Vitus Benson, Fanny Yang, Markus Reichstein），经 AITOP 多维评分管线结构化提炼。