| 0920-am-Su1-A-03-华为-光源集成需求 | Laser technologies for scale-up optical interconnects in AI supernodes | Scale Up |
| 0920-am-Su1-A-04-AMD-硅光片上光源 | Lasers for Co-packaged Optics（Light sources for next-generation optical communication systems for AI datacenters） | Scale Up |
| 0920-am-Su1-A-05-Columbia-片上光源与系统 | AI System Drivers for Laser Sources | Scale Up |
| 0920-am-Su2-A-03-PhotonBridge-集成激光器 | High-Power Lasers on Silicon: Scalable Multi-Wavelength Light Sources for CPO | Scale Up |
| 0920-pm-Su3+Su4-A-00-全场-A1下午全场扫描（第2–13页） | 未见完整英文原题；p3 议程标题为 “Navigating growth and uncertainty” | Scale Up |
| 0920-pm-Su3+Su4-A-00-全场-A1下午全场扫描（第14–23页） | AI [Cluster] Networks: Requirements for Reliable, High-Bandwidth-Density …（p14 标题 OCR 残缺，完整英文原题看不清） | Scale Up |
| 0920-pm-Su3+Su4-A-00-全场-A1下午全场扫描（第53–67页） | The Path for Co-Packaged Optics in the AI Ecosystem | Scale Up |
| 0920-pm-Su3+Su4-A-00-全场-A1下午全场扫描（第99–107页） | CPO, LPO, T/LRO, XPO, EIEIO…. A vendor perspective on architectural evolution | Scale Across |
| 0920-pm-Su3-A-01-LightCounting-AI连接硬件市场 | How Will AI Shape the Connectivity Market（副标题 Meeting Diverse AI Connectivity Needs: Architectural Choices for Next Generation Pluggable Transceivers） | Scale Up |
| 0920-pm-Su3-A-02-OpenAI-ScaleUp互连需求 | AI Scale-Up Networks: Requirements for Reliable, High-Bandwidth-Density Interconnects | Scale Up |
| 0920-pm-Su3-A-04-OrioleNetworks-数据中心光交换 | A Pure Photonic Network for the AI Era | Scale Up |
| 0920-pm-Su3-A-05-NVIDIA-CPO对可插拔的优势 | The Path for Co-Packaged Optics in the AI Ecosystem | Scale Up |
| 0920-am-Su1-B-01-Meta-AI数据中心功耗 | Standard Optic’s Perspective on Power Savings（首页标题；无独立总题页） | Scale Up |
| 0920-am-Su1-B-02-OpenAI-ScaleUp还能用铜吗 | AI scale-up – Does copper still do the job? | Scale Up |
| 0920-am-Su1-B-04-Broadcom-AI集群高速IO | High-Speed I/O in AI clusters | Scale Up |
| 0920-am-Su1-B-05-Arista-超密度可插拔XPO | Pluggable Forever: Power-Efficient, Ultra-Dense XPO — Breaking the AI Network Power & Density Wall | Scale Up |
| 0920-am-Su1-B-06-AttoTude-介质波导做ScaleUp | AttoTude Thesis – ASICs over Dielectrics（p2标题；DiAx介质增强双绞线） | Scale Up |
| 0920-am-Su2-B-02-Ciena-多轨光子破功耗墙 | Breaking the Power Wall with Multi-Rail Photonics and Optical Interconnect Innovation | Scale Across |
| 0920-pm-Su3-B-05-Lightera-铋掺放大器高功率（第1–11页） | Amplifiers for Hollow Core Fiber Transmission | Transport |
| 0920-pm-Su3-D-03-TUe-空分复用技术方（第1–5页） | SDM 技术视角（幻灯无独立英文题，随 Workshop 题：Heterogeneous optical networks） | Transport |
| 0920-pm-Su3-F-04-住友电工-数据中心光源用于PON | Leveraging Data-Center Optical Sources for VHSP PON: Opportunities for Coherent Lite and Hybrid Architectures | Access |
| 0920-pm-Su3-F-06-KIT-混合相干PON的挑战 | Challenges and Possibilities of Hybrid Coherent PONs | Access |
| 0920-pm-Su4-F-05-华为-直检的极限与变通 | Pushing Direct Detection: Limits and Workarounds for VHSP（Su3-F 场，“Hybrid Solutions… IMDD worlds?”，ECOC 2026, Malaga, 19 Sept） | Access |
| 0920-am-Su1-C-05-Tyndall-封装在产品化中的角色 | （封面题目文字OCR乱码，主题为封装在光子产品化中的角色；页2标题“Why is packaging so important?”） | Scale Up |
| 0920-pm-Su4-C-01-NTT-光网络数字孪生 | Towards AI-Native Infrastructure with Optical Network Digital Twins | Transport |
| 0920-pm-Su4-C-03-华为-算网一体设计 | Integrated Computing-Network Design for AIDC | Scale Across |
| 0920-am-Su2-I-01-NVIDIA-面向AI工厂的光子技术 | Photonics-enabled technologies for AI factories | Scale Up |
| 0920-am-Su2-I-03-iPronics-可扩展光子集成ScaleUp | Scalable Photonic Integration for energy-efficient scale-up AI networking | Scale Up |
| 0920-pm-Su3-I-04-Qualcomm-芯片侧的选择 | Beyond the SerDes: From Fast & Narrow to Reliable Connectivity at Scale | Scale Up |
| 0920-pm-Su4-I-02-OpenAI-ScaleUp需求 | Fast/Narrow or Slow/Wide? Choosing for Bandwidth Density and Reliability | Scale Up |
| 0920-pm-Su4-I-04-Coherent-器件路线 | （首页OCR为算力/互连受限图；未见明确英文题，内容围绕VCSEL阵列与DWDM的能耗/带宽比较） | Scale Up |
| 0920-pm-Su4-I-06-LightMatter-光计算路线 | 未见题目页（页1、2 OCR为空，未查看）；讲稿主题为 “Fast and narrow has a ceiling. Slow and wide does not — when it’s 3D.”（回答 Workshop Question：fast, narrow channel 能否无限上升） | Scale Up |
| 0920-pm-Su4-I-07-Credo-电互连路线 | 未见完整题目页（页1 OCR 残缺，未查看）；主题为 wide-parallel（宽并行）光互连，页标题 “When Wide-Parallel Could Bring Benefit” | Scale Up |
| 0921-MF-am-T02-1020-CignalAI-光电路交换的新应用 | New Applications for Optical Circuit Switching（据文件名；页内主标题看不清） | Scale Up |
| 0921-MF-am-T03-1040-Ciena-加速CPO的市场采用 | Accelerating Market Adoption of CPO（据文件名；页内 Keys for CPO/NPO ramp） | Scale Up |
| 0921-MF-am-T04-1100-华为-NPO与CPO的近期部署平衡 | Balancing NPO and CPO for near-term deployment（据文件名） | Scale Up |
| 0921-MF-am-T06-1140-Coherent-先进材料与散热效率 | Advanced materials and thermal efficiency（据文件名） | Scale Up |
| 0921-MF-am-T07-1200-Corning-先进光连接扩展AI互连-照片版（同场”非照片版”为同一讲的另一份拍摄，内容重复，合并记录） | Advanced Optical Connectivity to Scale AI Interconnect（据文件名；页内 “Context: CPO/NPO becoming a reality”） | Scale Up |
| 0921-MF-pm-1300-Arista-光互连演进-从可插拔到XPO与CPO | Optics interconnect evolution from pluggable to XPO and CPO（据文件名；标题页未看清） | Scale Up |
| 0921-MF-pm-1440-LIGENT-从可插拔到CPO-近封装光模块的角色 | From pluggable to CPO: the role of near-packaged optics（据文件名；标题页为空白） | Scale Up |
| 0921-MF-pm-1500-HeraeusCovantics-面向AI数据中心的下一代光纤 | Next-generation fiber technologies for AI DC（据页脚“…generation Fiber technologies for AI DC”，原题不完整） | Scale Up |
| 0921-Mo12-00-全场-开幕与大会报告全场连拍（第1–36页，Ciena大会报告） | 题目页OCR乱码，看不清；内容主线为 Scaling AI is all about scaling I/O（据p35结论页） | Scale Up |
| 0921-Mo12-00-全场-开幕与大会报告全场连拍（第85–107页，华为大会报告） | Scaling AI From Optical Innovation to NPO and CPO Solutions（p85） | Scale Up |
| 0921-Mo12-A3-Ciena-AI集群通信（共33页，单讲） | AI 集群通信主题的大会报告（英文原题页未拍到；末页标题 “Top Three Take-Aways”，主线为 Scaling AI is all about scaling I/O） | Scale Up |
| 0921-Mo12-A4-IMEC-集成光子支撑AI扩展（共25页，单讲） | 集成光子支撑 AI 扩展（英文原题页未看清；p1 标题页显示 “Future AI infrastructure is under tension”） | Scale Up |
| 0921-Mo12-A5-华为-从光创新到NPO与CPO（共14页，单讲） | 从光创新到 NPO 与 CPO（照片版原题页未清晰拍到；主线为 NPO vs CPO “多回合”辩论，结论 “NPO is the Optimal Solution for the 200G/Lane Era!”；最末页标题 “The Ubiquitous Optical Interconnect: Illuminating the Entire AI Network”） | Scale Up |
| 0921-Mo12-A6-PsiQuantum-光子量子计算（共34页，单讲） | 光子量子计算（英文原题页未拍到；p5 页 “How to build a scalable photonic quantum computer”） | 新应用 |
| 0921-Mo3-B1-领纤Linfiber-打破损耗壁垒-空芯光纤从物理极限到量产 | Breaking the Barrier: Hollow-Core Fibers from Physical Limits to Mass Production | Transport |
| 0921-Mo3-待定-Lightera-无中继海缆系统 | Real-Time 52.8 Tb/s (26.4 Tb/s/core) Unrepeatered Transmission over 303.4 km of Uncoupled 2-core MCF（ECOC2026 paper Mo3-G6） | Transport |
| 0921-Mo3-待定-Corning-AI互连的纵向与横向扩展路径 | Fiber Bundle Connector Solutions for VCSEL-Based Optical Scale-Up Interconnects | Scale Up |
| 0921-Mo3-待定-Tyndall-AI互连路径 | Glass Wafer-Level Packaging（含后续 Electro-Optical Interposer） | Scale Up |
| 0921-Mo4-待定-Arista-题目未公布（第1–24页；p10／p11／p16为另一版重复页） | 未公布（内容为 XPO 模块与 AI 互连五类场景，英文原题看不到） | Scale Up |
| 0921-Mo4-待定-Broadcom-为AI扩展光互连-走向大规模CPO集群（第1–17页；p9与p8内容相同） | Scaling Optical Interconnects for AI: Manufacturing Pathways to Massive CPO-Enabled Clusters | Scale Up |
| 0921-Mo4-待定-Cerebras-题目未公布（第1、9–16页；p2–p8为另一版重复页） | Heterogeneous Hybrid Bonding: A Path to Wafer-Scale Optical Systems | Scale Up |
| 0921-Mo4-待定-MicrosoftAzure-AI纵向扩展与通用算力的光互连用例（第1–10页） | 未见原题（文件名：AI 纵向扩展与通用算力的光互连用例；英文原题看不到） | Scale Up |
| 0921-Mo4-待定-NVIDIA-纵向与横向扩展的光方案（第1–12页） | Optical Solutions for Scale-out & Scale-up Connectivity | Scale Up |
| 0921-Mo5-待定-长飞YOFC-超高模纯度与单模匹配的O波段混合反谐振空芯光纤 | 题目页 OCR 乱码，按文件名与结论页为 O-band hybrid antiresonant HCF（HA-HCF）：ultrahigh modal purity and single-mode-matched（英文原题未完整核实） | Scale Across |
| 0921-Mo5-F5-NTT-168-GBd-PCS-324QAM与样本域Volterra预失真 | 168-GBd PCS-324QAM Signal Transmission with 2.1-Tb/s Net Bitrate Using Sample-Domain Volterra Nonlinear Pre-distortion（Mo5-F5） | Transport |
| 0921-PF-待定-Marvell-面向AI纵向扩展网络的近封装光模块 | Near Packaged Optics for AI Scale-up Networks | Scale Up |
| 0921-合集待拆-全场-Mo5-F全场连拍（第62–76页） | DPD-KAN: Kolmogorov-Arnold Networks for Low-Complexity Digital Predistortion in 5G Analog Radio-over-Fiber Systems | Access |
| 0921-合集待拆-全场-Mo5-G超高速PON连拍（第1–19页，Mo5-G1） | Low-Complexity VSB Generation and Chirp Management for VHSP Downstream Links Supporting GPON Coexistence | Access |
| 0922-MF-am-1000-Coherent-支撑AI基础设施三种扩展的光技术 | Photonics: from transport infrastructure to compute fabric（标题页看不清，据第2页大标题） | Scale Up |
| 0922-MF-am-1040-fibeReality-光器件厂商的高风险对冲 | 光器件厂商的高风险对冲（英文原题未见；含 “Verdict: Co-Packaged Optics Illusion Has Ended” 等观点页） | Scale Up |
| 0922-MF-am-1120-Jabil-面向大批量制造的光子封装 | 面向大批量制造的光子封装（英文原题标题页无 OCR/未看；内容为 photonics packaging for HVM） | Scale Up |
| 0922-MF-am-1220-EBOMSA-EBOMSA可靠且可互通的光互连 | EBO MSA: Enabling Reliable and Interoperable Optical Connectivity for AI Data Centers | Scale Up |
| 0922-MF-pm-1400-Broadcom-共封装光学之后再谈OCI | Co-Packaged Optics and Beyond: Optical Compute Interconnect (OCI)（英文原题据文件名推断，题目页未见清晰；p1 为拍摄的 “AI Interconnect” 页，OCR 全为乱码） | Scale Up |
| 0922-PF-0935-Advantest-硅光与共封装光学测试的量产化 | 未见完整原题（PDF名：硅光与共封装光学测试的量产化；内容为 SiPh/NPO/CPO HVM 测试） | Scale Up |
| 0922-PF-1155-SalienceLabs-把OCS从横向扩展带到纵向扩展 | 未见完整原题（PDF名：把OCS从横向扩展带到纵向扩展；页内出现 “OCS in scale-up networks”） | Scale Up |
| 0922-PF-1405-NewPhotonics-NPO互连作为CPO纵向扩展的务实策略 | The Architecture of AI Scale-Up: Why Near-Packaged Optics (NPO) is the Permanent Hybrid Standard | Scale Up |
| 0922-Tu1-C4-西湖大学-双偏振载波提取直检的448Gbps室内光无线接入 | Dual-Polarization Carrier-Extracted Direct Detection (DP-CEDD) for 448-Gbps Indoor Optical Wireless Access（题目页OCR乱码，按目录页”DP-CEDD / 448-Gbps / 9-m FSO link”整理，非原题逐字） | Access |
| 0922-Tu1-C5-KDDIResearch-光子晶体面发射激光器做自由空间光通信 | Free-Space Optical Communication using Photonic-Crystal Surface-Emitting Lasers | Access |
| 0922-Tu1-E2-浙江大学-2.5D环面拓扑的低损硅光MEMS开关阵列 | Low-loss silicon photonic MEMS switch array with 2.5D Torus topology（据小节标题概括，完整原题页看不清） | Scale Up |
| 0922-Tu1-E3-Microsoft-宽而慢架构加microLED打破AI网络与内存墙 | Wide-and-slow architecture with microLED to break AI network and memory walls（据文件名与内容概括，原题页未见） | Scale Up |
| 0922-Tu1-E5-NVIDIA-时钟前传DWDM光链路的微环静态与动态分配 | Static and Dynamic Ring Assignment in a Clock-Forwarded DWDM Optical Link | Scale Up |
| 0922-Tu1-G2-KIT-暗孤子微梳无源相干克隆的24波零差检测 | 24-λ Coherent Optical Communications Using Passive Coherence Cloning of Dark-Soliton Microcombs for Homodyne Detection（Tu1-G2） | Transport |
| 0922-Tu1-G5-Aston大学-172TbpsGMI的O波段相干传输 | 页面 p1 乱码，完整英文题看不清（主题为 172 Tb/s GMI、16.14 THz 满载 O 波段相干传输，Tu1-G5） | Transport |
| 0922-Tu3-A1-丹麦科技大学-AI时代的通信工程与收发机优化（第1–56页） | AI时代的通信工程与收发机优化（英文原题看不清，OCR乱码；会场 Tu3-A “AI and advanced processing for optical communication systems”） | Transport |
| 0922-Tu3-A3-NokiaBellLabs-用时域酉变换生成220GBd相干波形（第1–15页） | Generating 220 GBd coherent waveforms with spectro-temporal (time-domain) unitary transformations（英文原题页看不清，据结论/参考文献页推断；参考文献 J. Lightwave Technol. 44, 4880–4889 (2026)） | Transport |
| 0923-MF-00-上午连拍（第39–53页） | OIF: Developing AI Optical Interconnect Standards on the AI Timeline（页脚标注，23 Sep 2026） | Scale Up |
| 0923-MF-OIF-CMIS管理接口 | CMIS（Common Management Interface Specification）演进与增强固件管理（英文原题未见完整题目页，p1 OCR 乱码） | Scale Across |
| 0923-MF-euNetworks-市场聚焦 | p1 “Explosion of Solutions – We’ve been here before”（Cisco）；p2–p5 euNetworks 关于 AI 推理网络/小语言模型（SLM）的市场观点（题目页未拍到） | Transport |
| 0923-MF-康宁-市场聚焦 | Meeting the deployment challenges of optical scale-up | Scale Up |
| 0923-MF-美团-市场聚焦 | Rethinking Optical Interconnect for the AI Agent Era | Scale Up |
| 0923-MF-00-Marvell与Ciena与Arista与Oracle连拍（第33–40页） | Reality checking 1.6T+ AI interconnect solutions | Scale Up |
| 0923-MF-00-Marvell与Ciena与Arista与Oracle连拍（第41–45页） | Optics for NVIDIA Spectrum-X Multiplane Network Architecture | Scale Up |
| 0923-MF-00-Marvell与Ciena与Arista与Oracle连拍（第46–54页） | （标题页未见）内容涵盖 Optical connectivity expanding / Integrated optics requires a variety of technologies / Driving the future of pluggable transceivers | Scale Up |
| 0923-MF-00-Marvell与Ciena与Arista与Oracle连拍（第55–63页） | Open CPX 相关（Open CPX supporting broad NPO and CPO use cases / Diablo-1 6.4T Open CPX engine），标题栏被裁切 | Scale Up |
| 0923-We-F-00-标准化专场II连拍（第11–19页） | Optical Scale-Up AI Systems: progress in Standards | Scale Up |
| 0923-We1-G-00-全场连拍（第23–35页） | Highly Sensitive and Phase-Noise-Tolerant Multipath Interference Localization in IM/DD Systems by Block-based SNC | Transport |
| 0923-We1-H-00-全场连拍-PON的DSP与均衡（第68–84页） | Hybrid 50G DD and 200G Coherent PON with a Shared OLT based on 67-GHz TFLN MZM and Bias Optimization | Access |
| 0923-We2-I-00-全场连拍（第51–73页，第5讲） | Completely Optical-Filterless Comb-Based Transmitter with a Relaxed Clock Rate for Multi-Tbps Single-Carrier Transmission | Transport |
| 0923-We2-I-NTT-实时信号处理（第2–5、8–20页为非重复页） | Completely Optical-Filterless Comb-Based Transmitter with a Relaxed Clock Rate for Multi-Tbps Single-Carrier Transmission | Transport |
| 0923-We2-I-00-全场连拍（第74–84页，第6讲） | First-Ever Burst-Capable Adaptive Optical Signal Processor for Next-Generation PON Upstream Links | Access |
| 0923-We2-I-PCRL-实时信号处理（第5、6、8页为非重复页） | First-Ever Burst-Capable Adaptive Optical Signal Processor for Next-Generation PON Upstream Links | Access |
| 0923-We3-D3-KIT-光电任意波形生成 | Photonic-Electronic Arbitrary-Waveform Generation Enabling 276 GBd Electrical and 260 GBd Optical PAM4 Signaling | Transport |
| 0923-We3-D5-上海交大-高电谱效率直接检测（第1–25页；p7、13、14、16、18为重复） | Direct Detection with High Electrical Spectral Efficiency | Transport |
| 0923-We3-I-00-全场连拍-空芯表征与部署（第68–90页） | Field Deployment Practice, Key Challenges, and Standardization Outlook of Hollow Core Fibers and Transmission Systems | Transport |
| 0923-We5-B-Arista-光发射机与收发（第1–9页） | Industry’s First 12.8T 8×DR8 High-Density Liquid-Cooled 64-Channel Pluggable XPO Optical Module for Scale-Up | Scale Up |
| 0923-We5-B-NVIDIA-高速VCSEL与共封装（第1–31页，其中 p18／p20／p28 为重复页） | Progress of High-speed VCSELs and VCSEL-based Co-Packaging for Short Reach Communications | Scale Up |
| 0923-We5-G-00-全场连拍（第14–28页；重拍见 0923-We5-G-北京大学-相位与定时恢复） | Phase-Recovery-Free 6bit/4D Coherent Transmission Using Subcarrier-Joint Modulation for Large-Linewidth Lasers | Transport |
| 0923-We5-G-00-全场连拍（第42–52页；无单独 PDF） | Receiver-Side Compensation of TI-DAC Offset Mismatch in Optical Transceivers（We5-G4） | Transport |
| We5-I Coherent（第1–12页） | 2.3 Tbit/s Backside-Emitting 1060 nm VCSEL Array on Silicon Interposer for CPO Applications（Paper ID #504） | Scale Up |
| We5-I Lumentum（第1–18页） | 1060nm VCSEL Arrays for Scale-Up | Scale Up |
| We5-I 博升（第1–20页） | 1060 nm Back-Emitting VCSEL Modulated at 106 Gbps PAM4 Transmission over 100 m OM3 MMF for NPO Applications | Scale Up |
| We5-I 坦佩雷ORC 低温VCSEL（第1–19页） | Low-Threshold Single-Frequency Intra-Cavity VCSELs for Scalable Cryogenic Computing | 新应用 |
| 0924-PDP-A-1-Microsoft与南安普顿大学-C波段电信级损耗零色散空芯光纤首次验证 | First Demonstration of a Zero-Dispersion Hollow-Core Fibre with Telecoms-Grade Loss (0.2 dB/km), in the C-Band | Transport |
| 0924-PDP-A-2-MicrosoftAzureFiber-TFLN调制器与增益平坦YDFA及低损空芯光纤的1um相干传输系统 | A Coherent 1µm Transmission System using TFLN Modulator, Gain-Flattened YDFAs and Low-Loss Hollow-Core Fibre | Transport |
| 0924-PDP-A-5-NokiaBellLabs-450GBaud全电子IQ信号实现3.83Tbps单波长相干传输 | 3.83-Tb/s Single-Wavelength Coherent Transmission Enabled by 450-GBaud All-Electronic I/Q Signaling | Transport |
| 0924-PDP-C-5-NokiaBellLabs-双单边带传输实现340ps每nm色散容限的单波超速率200G直检PON下行 | Dual single-sideband transmission for 340 ps/nm CD tolerance, single-wavelength 200G direct-detection PON downstream（页面无完整英文题名，据文件名与内容推断；p1 OCR 乱码，未看图） | Access |
| 0924-Th1-D5-浙江大学-脊波导低损数字可调色散控制器（第1–19页；p14为p13重复，p20空白） | Low loss Digitally Tunable Dispersion Controller Using Ridge Waveguides | Transport |
| 0924-推定F2-Lightmatter-面向AI信息处理的光子学互连与内存带宽 | 标题页 OCR 乱码；按文件名为面向 AI 的光子互连与内存带宽，幻灯片主线为 Passage M1000 3D 光子”superchip”平台（英文原题看不清） | Scale Up |
| 0924-推定F3-哥伦比亚大学-面向AI集群的可编程光子学 | Programmable photonics for AI clusters | Scale Up |
| 0924-推定F4-iPronics-面向光交换的可编程光子学 | 页面无明确总题目；主题为面向 AI 数据中心的硅光可编程 OCS（Programmable photonics for optical switching） | Scale Up |
| 0924-Th1-G4-德国联邦国防军大学-可编程数据面与SDN实现任意业务承载于DWDM | Anything over Dense Wavelength Division Multiplexing with Programmable Data Planes and Software Defined Networking | Transport |