Simiao Ren
PulseAugur coverage of Simiao Ren — every cluster mentioning Simiao Ren across labs, papers, and developer communities, ranked by signal.
1 day(s) with sentiment data
-
Open-source AI image detectors show instability and struggle against modern generators
A new benchmark study evaluating 16 open-source AI-generated image detection models has revealed significant performance gaps and instability across various datasets. The research, which analyzed 2.6 million images from…
-
Deepfake detectors fail on real-world videos due to post-processing
A new study published on arXiv reveals that common post-processing techniques, specifically super-resolution imaging, significantly degrade the effectiveness of existing deepfake detection methods. Researchers collected…
-
AI-generated loan document forgeries detected by human reviewers in study
A new study published on arXiv explores the ability of human loan processors and underwriters to detect AI-generated forgeries in financial documents. Researchers created 134 forged PDFs from genuine financial statement…
-
AI agents can forge PDF documents with high success rate, study finds
A new benchmark, AgentForge-Bench, has been developed to measure the ability of AI agents to forge PDF documents. The benchmark found that off-the-shelf coding agents, utilizing open-weight models, can successfully alte…
-
Synthetic eye movement dataset generated for AI research
Researchers have developed a novel pipeline for generating synthetic eye movement data, addressing the scarcity and privacy concerns associated with real-world behavioral data. This method involves extracting human iris…
-
New benchmark reveals critical calibration failures in document forgery detection
A new benchmark called DOCFORGE-BENCH has been introduced to evaluate document forgery detection methods in a zero-shot setting, meaning without domain-specific training data. The benchmark found that current methods su…
-
AI age estimators vulnerable to simple cosmetic changes, study finds
A new research paper explores the vulnerability of AI age estimation systems to low-cost cosmetic attacks. The study simulated modifications like beards, grey hair, makeup, and wrinkles on facial images of individuals a…
-
Multi-modal LLMs evaluated for document fraud detection capabilities
A new study published on arXiv explores the capabilities of various multi-modal large language models (LLMs) in detecting document manipulation. Researchers benchmarked models such as OpenAI's O1 and 4o, Gemini Flash, D…
-
Small language models boosted for AI call screening with new training data
Researchers have developed a method to improve small language models (SLMs) for AI call-screening agents by creating a quality-controlled dataset. This dataset, comprising 3,395 synthetic calls, focuses on generating re…
-
Multi-modal LLMs tested for deepfake detection capabilities
A new research paper explores the effectiveness of multi-modal large language models (LLMs) in detecting deepfakes. The study benchmarks 12 LLMs, including models like OpenAI O1/4o and Gemini thinking Flash 2, against t…
-
Humans struggle to detect AI-generated receipts, LLMs excel
A new dataset called GPT4o-Receipt has been released to study the forensics of AI-generated financial documents. The dataset includes 1,235 receipt images, comparing those generated by GPT-4o with authentic receipts. Hu…
-
New benchmark CallScreenBench evaluates small language models as phone secretaries
Researchers have introduced CallScreenBench, a new benchmark designed to evaluate the performance of small language models acting as phone secretaries. This benchmark focuses on the conversational decision-making layer,…