TRIM: Reducing AI-Generated CodeSlop via Agent Trajectory Minimization
3 items
- The Inference Weekly Trim the Slop: How Trajectory-Guided Redundancy Minimization Makes AI-Generated Code Cleaner
- Scan AI Slop CodeSlop: Why AI Coding Agents Leave Abandoned Edits Behind
- AgentPatterns.ai CodeSlop: Search-Trajectory Residue in Agent Patches
Outrunning LLM Cutoffs: A Live Kernel Crash Resolution Benchmark for All
1 item
- Columbia DAPLab Columbia DAPLab at ICML 2026
zkFuzz: Foundation and Framework for Effective Fuzzing of Zero-Knowledge Circuits
1 item
- Columbia Computer Science Security Research Recognized at IEEE S&P 2026
Your Compiler is Backdooring Your Model: Understanding and Exploiting Compilation Inconsistency Vulnerabilities in Deep Learning Compilers
3 items
- Columbia Computer Science Security Research Recognized at IEEE S&P 2026
- The Weather Report Your Compiler is Backdooring Your Model
- The Weather Report Predicting AI Attacks from IEEE S&P 2026 Papers (Preview)
Radshield: Software Radiation Protection for Commodity Hardware in Space
1 item
- Dangling Pointers Radshield: Software Radiation Protection for Commodity Hardware in Space
PickleBall: Secure Deserialization of Pickle-based Machine Learning Models
1 item
- James Davis on Medium PickleBall: Secure Deserialization of Pickle-based Machine Learning Models
EditLord: Learning Code Transformation Rules for Code Editing
1 item
- Chameleon Cloud EditLord: Learning Code Transformation Rules for Code Editing
Do Spammers Dream of Electric Sheep? Characterizing the Prevalence of LLM-Generated Malicious Emails
16 items
- Forbes AI Is Behind 50% of Spam — And Now It’s Hacking Your Accounts
- TechRepublic AI Now Creates 51% of Spam: Two Key Reasons Attackers Use This Approach
- Infosecurity Magazine AI Now Generates Majority of Spam and Malicious Emails
- The Register AI Improves Spam
- Columbia Engineering AI Now Powers Over Half of Spam Emails, Columbia Engineering Research Finds
- Columbia Magazine The Deepfake Scam Era Is Upon Us. Here’s How to Get Ready.
- Barracuda Half the Spam in Your Inbox Is Generated by AI — Its Use in Advanced Attacks Is at an Earlier Stage
- TechRadar Pro Major Spam Email Warning — AI Now Generates Almost All of Your Junk Mail
- Cybernews Scammers Now Using AI for A/B Testing
- Security Boulevard Analysis Surfaces Increased Usage of LLMs to Craft BEC Attacks
- Channel Insider AI Now Powers Over Half of Global Spam Emails, Research Finds
- Cloudflare Attackers Go Phishing in New Ponds
- DIGIT Over Half of All Spam Emails Are Now AI-Generated
- CybersecAsia Fraudsters and Cybercriminals Tap AI for More Sophisticated Spam and BEC Attacks
- SecurityBrief Australia Over Half of Spam Emails Now Generated by AI, New Study Finds
- ANSA Il 51% delle email indesiderate è generato dall'IA
Raidar: geneRative AI Detection viA Rewriting
8 items
- Science News Explores A New Test Could Help Weed Out AI-Generated Text
- Columbia Engineering Who Wrote This? Columbia Engineers Discover Novel Method to Identify AI-Generated Text
- Columbia Magazine A New Way to Spot Text Written by AI, and Other Science News
- Columbia Engineering Turns Out, I’m Not Real: Detecting AI-Generated Videos
- 36Kr SORA、Gen-2、Pika也逃不过，文生视频检测新工具来了，准确率高达93.7%
- Science News Explores Google Now Adds Watermarks to All Its AI-Generated Content
- Tech Xplore Who Wrote This? Engineers Discover Novel Method to Identify AI-Generated Text
- Tech Xplore New Tool Detects AI-Generated Videos with 93.7% Accuracy
Chablis: Fast and General Transactions in Geo-Distributed Systems
2 items
QUACK: Hindering Deserialization Attacks via Static Duck Typing
1 item
Turns Out I'm Not Real: Towards Robust Detection of AI-Generated Videos
7 items
- Columbia Engineering Turns Out, I’m Not Real: Detecting AI-Generated Videos
- TVNewsCheck Is That an AI Video? Here’s How Newsrooms Can Spot Them
- 36Kr SORA、Gen-2、Pika也逃不过，文生视频检测新工具来了，准确率高达93.7%
- VIDI Otkrivanje AI generiranog videa s velikom točnošću
- Vetenskapsnytt Nytt verktyg upptäcker AI-genererade videor med 93,7% noggrannhet
- Tech Xplore New Tool Detects AI-Generated Videos with 93.7% Accuracy
- EurekAlert! Turns Out I’m Not Real: Detecting AI-Generated Videos
Chardonnay: Fast and General Datacenter Transactions for On-Disk Databases
2 items
Divergent Representations: When Compiler Optimizations Enable Exploitation
1 item
- Trail of Bits Look Out! Divergent Representations Are Everywhere!
UPGRADVISOR: Early Adopting Dependency Updates Using Hybrid Program Analysis and Hardware Tracing
1 item
- Columbia Computer Science Software Systems Laboratory Wins Best Paper Award at OSDI 2022
XRP: In-Kernel Storage Functions with eBPF
5 items
- Columbia Electrical Engineering Prof Asaf Cidon and Team Receive Jay Lepreau Best Paper Award at 16th USENIX Symposium on Operating Systems Design and Implementation
- Columbia Computer Science Software Systems Laboratory Wins Best Paper Award at OSDI 2022
- HeapDump XRP：用eBPF优化内存存储功能
- Hacker News XRP: In-Kernel Storage Functions with eBPF
- BBSakura Networks eBPF の個人的おもしろトピック
Argus: Debugging Performance Issues in Modern Desktop Applications with Annotated Causal Tracing
1 item
- Columbia Computer Science Software Systems Laboratory Wins Best Paper Awards at the OSDI and USENIX ATC Conferences
BPF for Storage: An Exokernel-Inspired Approach
4 items
- eBPF Reading Group BPF for Storage: An Exokernel-Inspired Approach
- Hacker News BPF for Storage: An Exokernel-Inspired Approach
- Lobsters BPF for Storage: An Exokernel-Inspired Approach
- 博客园 BPF for storage：一种受外核启发的反式
Multitask Learning Strengthens Adversarial Robustness
2 items
- TOPBOTS ECCV 2020: Some Highlights
- LeiPhone 【重磅整理】ECCV 2020 亮点摘要
Towards Practical Verification of Machine Learning: The Case of Computer Vision Systems
1 item
- Columbia Engineering Magazine Machine Learning 2.0
Egalito: Layout-Agnostic Binary Recompilation
1 item
- Trail of Bits Towards Practical Security Optimizations for Binaries
DeepXplore: Automated Whitebox Testing of Deep Learning Systems
20 items
- Scientific American When AI Steers Us Astray
- IEEE Spectrum A New Way to Find Bugs in Self-Driving AI Could Save Lives
- Newsweek Scientists May Have Found a Way to Stop Artificial Intelligence from Becoming Racist and Sexist
- Communications of the ACM CACM Nov. 2019 - DeepXplore: Automated Whitebox Testing of Deep Learning Systems
- TechRadar Self-Driving Cars: Your Complete Guide to Autonomous Vehicles
- The Next Web Science May Have Cured Biased AI
- Columbia News Researchers Unveil Tool to Debug ‘Black Box’ Deep Learning Algorithms
- The Morning Paper DeepXplore: Automated Whitebox Testing of Deep Learning Systems
- The Foretellix CTO Blog DeepXplore and New Ideas for Verifying ML Systems
- EurekAlert First White-Box Testing Model Finds Thousands of Errors in Self-Driving Cars
- EurekAlert Researchers Unveil Tool to Debug ‘Black Box’ Deep Learning Algorithms
- Columbia Engineering Magazine Machine Learning 2.0
- LeiPhone 帮助自动驾驶找出算法中的 Bug，这种方法可堪大任吗？
- ChainNews 学界 | 新研究提出DeepXplore：首个系统性测试现实深度学习系统的白箱框架
- Sohu DeepXplore最新动态：科学有望打开AI“黑盒子”，消除AI偏见
- Sina 深度 | 详解首个系统性测试现实深度学习系统的白箱框架DeepXplore
- CCTV Hello AI Documentary — DeepXplore Segment
- e15.cz DeepXplore hledá chyby a nebezpečí v neznámých útrobách hlubokých neuronových sítí
- N+1 Тестирование нейросетей указало на дефекты в «мышлении» беспилотных автомобилей
- Columbia Engineering Luca Carloni and Junfeng Yang Elected 2025 ACM Fellows
NEUZZ: Efficient Fuzzing with Neural Program Smoothing
1 item
Efficient Repair of Polluted Machine Learning Systems via Causal Unlearning
1 item
Overload Control for Scaling WeChat Microservices
3 items
- The Morning Paper Overload Control for Scaling WeChat Microservices
- Tencent Cloud Developer Community 国民级应用：微信是如何防止崩溃的？
- Tushar Tyagi Overload Control for Scaling WeChat Microservices
Shuffler: Fast and Deployable Continuous Code Re-Randomization
5 items
- Network World Solution to JIT-ROP Cyber Attacks: Scramble Code Quickly
- ACM TechNews New Software Continuously Scrambles Code to Foil Cyberattacks
- Columbia Science Commits New Software Continuously Scrambles Code to Foil Cyber Attacks
- Digital Journal New Software Continuously Scrambles Code to Foil Cyber Attacks
- TechXplore New Software Continuously Scrambles Code to Foil Cyber Attacks
Reducing Crash Recoverability to Reachability
1 item
- The Morning Paper Reducing Crash Recoverability to Reachability
Towards Making Systems Forget with Machine Unlearning
12 items
- The Atlantic Teaching a Computer to Forget
- The Stack Machine Unlearning: How Can Information Be Forgotten in the Age of Viral Data Spread?
- KurzweilAI New Machine Unlearning Technique Deletes Unwanted Data
- ACM TechNews Machine Unlearning: How Can Information Be Forgotten in the Age of Viral Data Spread?
- EurekAlert New Machine Unlearning Technique Wipes Out Unwanted Data Quickly and Completely
- Lehigh News Machine Unlearning
- Lehigh Engineering Helping Machines Forget
- Columbia Computer Science Four Professors Recognized with Test of Time Awards
- Johns Hopkins Information Security Institute Yinzhi Cao Receives Distinguished Paper Award and Test-of-Time Award at IEEE Security and Privacy 2025
- IEEE Cipher Security and Privacy Symposium Test of Time Awards
- Columbia Engineering Magazine Machine Learning 2.0
- Columbia Engineering Luca Carloni and Junfeng Yang Elected 2025 ACM Fellows
Efficient Deterministic Multithreading through Schedule Relaxation
5 items
- The Register Better Multithreading Offered by Columbia U Researchers
- ACM TechNews Major Breakthrough Improves Software Reliability and Security
- Columbia News Major Breakthrough Improves Software Reliability and Security
- Phys.org Major Breakthrough Improves Software Reliability and Security
- TG Daily New System Could Improve Software Reliability and Security
MODIST: Transparent Model Checking of Unmodified Distributed Systems
2 items
EXPLODE: a Lightweight, General System for Finding Serious Storage System Errors
2 items
Automatically Generating Malicious Disks using Symbolic Execution
1 item
- LWN.net KHB: Automating Bug Hunting