MARL detector repairs adversarial attacks in energy forecasting
MARL detector and signal repair achieve F1 score above 0.91 and reduce error by up to 83% with low latency.
Stories per day · 30 days
Relatedadversarial-training1ai-security1black-box1concept-based1deep-learning1energy-forecasting1generalization1hiddenlayer1image-classifiers1jev1marl1neural-networks1occlusion1prompt-injection1query-efficiency1robustness1system-one1tokenbreak1
12 stories, newest first · one card per story
Stories tagged #adversarial-attacks
Dissertation explores bad minima in transfer learning via adversarial attacks
Adversarial attacks on transfer learning reveal a strong link between source generalization and successful transfer.
Concept-based attacks target neural-network logic, not class labels
Retraining on the attacks kept classical robustness gains while better removing semantic shortcuts.
STROKE uses sparse rectangular occlusions for black-box image attacks
The score-based method reports high success at 1–2% area with low query counts and shorter wall-clock time than matched patch attacks.
TypeSafe’s Jev ‘even more susceptible to attack’ than LLMs, says security vendor
HiddenLayer finds TypeSafe's Jev 'System One' model vulnerable to TokenBreak and prompt injection attacks that traditional LLMs typically resist.
Botnets, adversarial attacks and data poisoning top leaders’ AI threat list
PwC survey finds leaders least prepared for AI-directed botnets, adversarial attacks, and data poisoning.
3
MADBench: Benchmarking the Security of Multi-Agent Debate
MADBench shows multi-agent debate can reduce some answer attacks but amplify unauthorized reads and writes.
SoK: Trading Agents or Market Crashers? Dissecting Robustness and Security Failures in Academic Financial LLM Trading Schemes
FARSIGHT framework finds 80% of 15 academic financial LLM trading schemes fail robustness metrics and 100% exhibit security vulnerabilities.
Epsilon-Nash Equilibria in History-Dependent SA-MDPs
Researchers give the first algorithm for computing epsilon-approximate history-dependent equilibria in state-adversarial Markov decision processes with observation-perturbing adversaries.
MarkSec: Capability-Aware Evaluation of Adversarial Attacks Against LLM Watermarks
MarkSec unifies evaluation of stealing, scrubbing, and spoofing attacks against LLM watermarks with quality-constrained success metrics under shared reporting protocols.
DRIFT: Removing Diffusion Watermarks by Deflecting the Generative Trajectory
Introduces DRIFT, a black-box attack removing diffusion watermarks by deflecting generative trajectories, achieving 98-100% success across nine watermarking schemes.
Rethinking Indirect Prompt Injection as a Test-Time Search Problem
Researchers frame indirect prompt injection as test-time search, showing added attacker compute improves vulnerability discovery and exploitation against tool-using agents.