diff --git a/cache/seen_urls.json b/cache/seen_urls.json
index 2db1eb7..45e560f 100644
--- a/cache/seen_urls.json
+++ b/cache/seen_urls.json
@@ -1,5 +1,5 @@
 {
-  "lastUpdate": "2026-02-28T15:00:39.694Z",
+  "lastUpdate": "2026-03-03T15:00:57.517Z",
   "urls": [
     "http://arxiv.org/abs/2602.23360v1",
     "http://arxiv.org/abs/2602.23359v1",
@@ -143,6 +143,136 @@
     "https://github.com/login?return_to=%2fpaddlepaddle%2fpaddle",
     "https://github.com/login?return_to=%2fdatagouv%2fdatagouv-mcp",
     "https://github.com/login?return_to=%2fwei-shaw%2fsub2api",
-    "https://github.com/login?return_to=%2fx-plug%2fmobileagent"
+    "https://github.com/login?return_to=%2fx-plug%2fmobileagent",
+    "https://github.com/login?return_to=%2fmicrosoft%2fmarkitdown",
+    "https://github.com/login?return_to=%2fk-dense-ai%2fclaude-scientific-skills",
+    "https://github.com/login?return_to=%2fbasecamp%2fomarchy",
+    "https://github.com/login?return_to=%2fnevamind-ai%2fmemu",
+    "https://huggingface.co/papers/2602.21877",
+    "https://huggingface.co/papers/2602.23047",
+    "https://huggingface.co/papers/2602.24264",
+    "https://huggingface.co/papers/2602.23881",
+    "https://huggingface.co/papers/2602.20913",
+    "https://huggingface.co/papers/2602.24012",
+    "https://huggingface.co/papers/2602.22207",
+    "https://huggingface.co/papers/2602.24289",
+    "https://huggingface.co/papers/2602.23452",
+    "https://huggingface.co/papers/2602.22661",
+    "https://huggingface.co/papers/2602.23898",
+    "https://huggingface.co/papers/2602.24281",
+    "https://huggingface.co/papers/2602.24286",
+    "https://huggingface.co/papers/2602.24233",
+    "https://huggingface.co/papers/2602.23622",
+    "https://huggingface.co/papers/2602.23996",
+    "https://huggingface.co/papers/2602.24208",
+    "https://huggingface.co/papers/2602.22647",
+    "https://github.com/login?return_to=%2fanthropics%2fprompt-eng-interactive-tutorial",
+    "https://github.com/sponsors/servo",
+    "http://arxiv.org/abs/2603.02203v1",
+    "http://arxiv.org/abs/2603.02200v1",
+    "http://arxiv.org/abs/2603.02196v1",
+    "http://arxiv.org/abs/2603.02193v1",
+    "http://arxiv.org/abs/2603.02190v1",
+    "http://arxiv.org/abs/2603.02184v1",
+    "http://arxiv.org/abs/2603.02181v1",
+    "http://arxiv.org/abs/2603.02178v1",
+    "http://arxiv.org/abs/2603.02175v1",
+    "http://arxiv.org/abs/2603.02170v1",
+    "http://arxiv.org/abs/2603.02156v1",
+    "http://arxiv.org/abs/2603.02155v1",
+    "http://arxiv.org/abs/2603.02154v1",
+    "http://arxiv.org/abs/2603.02153v1",
+    "http://arxiv.org/abs/2603.02150v1",
+    "http://arxiv.org/abs/2603.02129v1",
+    "http://arxiv.org/abs/2603.02128v1",
+    "http://arxiv.org/abs/2603.02123v1",
+    "http://arxiv.org/abs/2603.02119v1",
+    "http://arxiv.org/abs/2603.02115v1",
+    "http://arxiv.org/abs/2603.02096v1",
+    "http://arxiv.org/abs/2603.02095v1",
+    "http://arxiv.org/abs/2603.02091v1",
+    "http://arxiv.org/abs/2603.02087v1",
+    "http://arxiv.org/abs/2603.02081v1",
+    "http://arxiv.org/abs/2603.02072v1",
+    "http://arxiv.org/abs/2603.02070v1",
+    "http://arxiv.org/abs/2603.02069v1",
+    "http://arxiv.org/abs/2603.02062v1",
+    "http://arxiv.org/abs/2603.02056v1",
+    "http://arxiv.org/abs/2603.02050v1",
+    "http://arxiv.org/abs/2603.02041v1",
+    "http://arxiv.org/abs/2603.02029v1",
+    "http://arxiv.org/abs/2603.02025v1",
+    "http://arxiv.org/abs/2603.02208v1",
+    "http://arxiv.org/abs/2603.02176v1",
+    "http://arxiv.org/abs/2603.02146v1",
+    "http://arxiv.org/abs/2603.02112v1",
+    "http://arxiv.org/abs/2603.02099v1",
+    "http://arxiv.org/abs/2603.02098v1",
+    "http://arxiv.org/abs/2603.02097v1",
+    "http://arxiv.org/abs/2603.02084v1",
+    "http://arxiv.org/abs/2603.02082v1",
+    "http://arxiv.org/abs/2603.02026v1",
+    "http://arxiv.org/abs/2603.02024v1",
+    "http://arxiv.org/abs/2603.02023v1",
+    "http://arxiv.org/abs/2603.01990v1",
+    "http://arxiv.org/abs/2603.01973v1",
+    "http://arxiv.org/abs/2603.01966v1",
+    "http://arxiv.org/abs/2603.01950v1",
+    "http://arxiv.org/abs/2603.01945v1",
+    "http://arxiv.org/abs/2603.01930v1",
+    "http://arxiv.org/abs/2603.01914v1",
+    "http://arxiv.org/abs/2603.01912v1",
+    "http://arxiv.org/abs/2603.01910v1",
+    "http://arxiv.org/abs/2603.01907v1",
+    "http://arxiv.org/abs/2603.01875v1",
+    "http://arxiv.org/abs/2603.01869v1",
+    "http://arxiv.org/abs/2603.01865v1",
+    "http://arxiv.org/abs/2603.02204v1",
+    "http://arxiv.org/abs/2603.02202v1",
+    "http://arxiv.org/abs/2603.02194v1",
+    "http://arxiv.org/abs/2603.02188v1",
+    "http://arxiv.org/abs/2603.02174v1",
+    "http://arxiv.org/abs/2603.02159v1",
+    "http://arxiv.org/abs/2603.02145v1",
+    "http://arxiv.org/abs/2603.02142v1",
+    "http://arxiv.org/abs/2603.02109v1",
+    "http://arxiv.org/abs/2603.02100v1",
+    "http://arxiv.org/abs/2603.02092v1",
+    "https://huggingface.co/papers/2603.01096",
+    "https://huggingface.co/papers/2603.01493",
+    "https://huggingface.co/papers/2602.23543",
+    "https://huggingface.co/papers/2603.00516",
+    "https://huggingface.co/papers/2603.01907",
+    "https://huggingface.co/papers/2603.00961",
+    "https://huggingface.co/papers/2602.19896",
+    "https://huggingface.co/papers/2603.01824",
+    "https://huggingface.co/papers/2602.23866",
+    "https://huggingface.co/papers/2603.01281",
+    "https://huggingface.co/papers/2603.02049",
+    "https://huggingface.co/papers/2603.02208",
+    "https://huggingface.co/papers/2603.00141",
+    "https://huggingface.co/papers/2603.00889",
+    "https://huggingface.co/papers/2602.23359",
+    "https://huggingface.co/papers/2603.00585",
+    "https://huggingface.co/papers/2603.02024",
+    "https://huggingface.co/papers/2603.01940",
+    "https://huggingface.co/papers/2603.01142",
+    "https://huggingface.co/papers/2603.02099",
+    "https://huggingface.co/papers/2603.01223",
+    "https://huggingface.co/papers/2603.02203",
+    "https://huggingface.co/papers/2603.01301",
+    "https://huggingface.co/papers/2603.02138",
+    "https://huggingface.co/papers/2603.01896",
+    "https://huggingface.co/papers/2603.01425",
+    "https://huggingface.co/papers/2603.01840",
+    "https://huggingface.co/papers/2603.00483",
+    "https://huggingface.co/papers/2603.01973",
+    "https://huggingface.co/papers/2603.01562",
+    "https://github.com/login?return_to=%2fruvnet%2fruview",
+    "https://github.com/login?return_to=%2fcodebuffai%2fcodebuff",
+    "https://github.com/login?return_to=%2fagentscope-ai%2fagentscope",
+    "https://github.com/login?return_to=%2fagentscope-ai%2freme",
+    "https://github.com/login?return_to=%2flmcache%2flmcache",
+    "https://github.com/login?return_to=%2faquasecurity%2ftrivy"
   ]
 }
\ No newline at end of file
diff --git a/daily/2026-03-03_en.md b/daily/2026-03-03_en.md
new file mode 100644
index 0000000..70a826d
--- /dev/null
+++ b/daily/2026-03-03_en.md
@@ -0,0 +1,50 @@
+# AI Daily Brief - 2026-03-03
+
+> Collected at: 3/3/2026, 11:00:57 PM
+> Total items: 106
+
+## 🔥 Top 10 Highlights
+
+1. From Scale to Speed: Adaptive Test-Time Scaling for Image Editing
+   https://huggingface.co/papers/2603.00141
+
+2. login?return_to=%2Fruvnet%2FRuView
+   https://github.com/login?return_to=%2Fruvnet%2FRuView
+
+3. login?return_to=%2FCodebuffAI%2Fcodebuff
+   https://github.com/login?return_to=%2FCodebuffAI%2Fcodebuff
+
+4. login?return_to=%2FLMCache%2FLMCache
+   https://github.com/login?return_to=%2FLMCache%2FLMCache
+
+5. login?return_to=%2Faquasecurity%2Ftrivy
+   https://github.com/login?return_to=%2Faquasecurity%2Ftrivy
+
+6. SageBwd: A Trainable Low-bit Attention
+   http://arxiv.org/abs/2603.02170v1
+
+7. Boltzmann-based Exploration for Robust Decentralized Multi-Agent Planning
+   http://arxiv.org/abs/2603.02154v1
+
+8. PonderLM-3: Adaptive Token-Wise Pondering with Differentiable Masking
+   http://arxiv.org/abs/2603.02023v1
+
+9. AdaPonderLM: Gated Pondering Language Models with Token-Wise Adaptive Depth
+   http://arxiv.org/abs/2603.01914v1
+
+10. Partial Causal Structure Learning for Valid Selective Conformal Inference under Interventions
+   http://arxiv.org/abs/2603.02204v1
+
+## 📂 Categories
+
+### AI Infrastructure / Inference Optimization
+
+- Multi-Head Low-Rank Attention
+  http://arxiv.org/abs/2603.02188v1
+- Instrumental and Proximal Causal Inference with Gaussian Processes
+  http://arxiv.org/abs/2603.02159v1
+- Orchestrating Multimodal DNN Workloads in Wireless Neural Processing
+  http://arxiv.org/abs/2603.02109v1
+
+---
+*Generated by AINewsCollector*
diff --git a/daily/2026-03-03_zh.md b/daily/2026-03-03_zh.md
new file mode 100644
index 0000000..3b8cf17
--- /dev/null
+++ b/daily/2026-03-03_zh.md
@@ -0,0 +1,50 @@
+# AI Daily Brief - 2026-03-03
+
+> 采集时间: 2026/3/3 23:00:30
+> 总条目: 106
+
+## 🔥 Top 10 重要消息
+
+1. 自适应测试时缩放：图像编辑速度提升新范式
+   https://huggingface.co/papers/2603.00141
+
+2. WiFi DensePose：将WiFi信号转化为实时人体姿态、生命体征监测与存在检测
+   https://github.com/login?return_to=%2Fruvnet%2FRuView
+
+3. 从终端生成代码的AI助手
+   https://github.com/login?return_to=%2FCodebuffAI%2Fcodebuff
+
+4. 加速大语言模型：打造最快KV缓存层
+   https://github.com/login?return_to=%2FLMCache%2FLMCache
+
+5. 容器安全检测：Trivy全面扫描漏洞与配置
+   https://github.com/login?return_to=%2Faquasecurity%2Ftrivy
+
+6. SageBwd：可训练低比特注意力加速模型训练
+   http://arxiv.org/abs/2603.02170v1
+
+7. 基于玻尔兹曼探索的鲁棒去中心化多智能体规划
+   http://arxiv.org/abs/2603.02154v1
+
+8. PonderLM-3：自适应分词级Pondering与可微分掩码技术
+   http://arxiv.org/abs/2603.02023v1
+
+9. AdaPonderLM：基于自适应深度的门控Ponder语言模型
+   http://arxiv.org/abs/2603.01914v1
+
+10. 干预下部分因果结构学习实现有效选择性一致预测
+   http://arxiv.org/abs/2603.02204v1
+
+## 📂 分类汇总
+
+### AI 基础设施 / 推理优化
+
+- 多头低秩注意力加速长文本推理
+  http://arxiv.org/abs/2603.02188v1
+- 基于高斯过程的工具变量与邻近因果推断方法研究
+  http://arxiv.org/abs/2603.02159v1
+- 无线神经处理中多模态DNN工作负载协同编排
+  http://arxiv.org/abs/2603.02109v1
+
+---
+*Generated by AINewsCollector*