Register and share your invite link to earn from video plays and referrals.

Search results for SearchAgents
SearchAgents community
One keyword maps to one global community path.
Create community
People
Not Found
Tweets including SearchAgents
TL;DR: When attacks on search agents evolve to adapt to the defenses, they cut accuracy far more than fixed attacks. Static benchmarks may be underestimating the risk. The Web Doesn't Sit Still: Adversarial Self-Evolving Attacks on Search Agents Key points ・🧪 Evaluates RAG poisoning with evolving attacks instead of fixed templates ・🧬 CREO: contrasts successful and failed reasoning traces, and an LLM crosses over and mutates strategies ・🔁 A bi-level setup where the defender discloses last round's attack patterns in its prompt ・📉 GLM-5.1 on MuSiQue: base Acc 82 drops to 41 under AuthChain and 16 under CREO ・⚠️ Mean TASR is 60.5–74.6% for CREO versus about 20% for FactSwap ・😵 A new metric, APR, also measures "cognitive paralysis", where agents refuse to answer after confusing contradictions ・🛡️ Keyword and perplexity filters barely help; even the effective defense leaves large residual risk The evaluation covers multi-hop QA only, and the defender knows the strategies. Still, it suggests search-agent safety evaluation should assume adaptive attackers. #AISecurity# #SearchAgents#
Show more
🔎 A paper that pinpoints the hidden reason RL training for search agents stalls partway. Title: Harness-G: A Graph-Structured Harness for Search Agents URL: ❓ Why does training collapse? 💡 It's "retrieval-equivalence collapse." The policy keeps generating differently-worded queries that fetch the same evidence, so same evidence → same answer → same reward, within-group advantages vanish, and the training signal dries up. ❓ How does Harness-G fix it? 💡 It stops free-form query generation and turns it into menu selection over a paragraph-sentence-entity graph built from the corpus. The policy picks action IDs, not strings. Being finite, verifiable, and previewable, it preserves diversity in what actually gets retrieved. ❓ What is the credit assignment (SNC)? 💡 A frozen answerer previews how much an action raises the gold-answer probability, scored against alternatives (frontier-relative). Non-myopic payoffs like "find the bridge entity first" propagate back through provenance edges (enablement). No extra rollouts needed. ❓ Does it work? 💡 Across six QA benchmarks it beats Graph-R1 by +10.74 at 1.5B and +3.98 at 3B, best at both scales, especially on multi-hop, with $0 API cost to build the graph. #SearchAgents# #ReinforcementLearning#
Show more
Even the top search agents set to the highest effort modes cannot find 1/3 of the verified information. Read the full research:
🔄 TL;DR: A search index that diagnoses its own weaknesses, rewrites its keys, and validates the changes — all without any human in the loop. It beats existing methods by a wide margin on BRIGHT. Title: Self-Evolving Search Index URL: Points 🩺 Builds co-retrieval profiles from search results to self-diagnose whether documents are well distinguished, with no human annotation needed ✍️ Selectively revises key sets only for flagged documents, autonomously deciding up to 10 keys per document ✅ Self-validates proposals on faithfulness, specificity, and separation, keeping only keys that pass 🔍 A Query Simulator proactively probes uncovered demand with synthetic queries 📈 Hits 22.8 average nDCG@10 on BRIGHT, +9.2% over RL-Index and +40-57% over the base index 🤖 For search agents, answer accuracy jumps +77.87% while search calls drop -16.80% The idea of making index optimization itself self-evolving, not just the model or the agent, is what makes this interesting. #InformationRetrieval# #LLMAgents#
Show more
Speed is key for @juicebox_work. They need to search hundreds of thousands of talent profiles in seconds with multiple real-time search agents. We helped them cut latency 80% and drop inference costs from $4M to $800K/yr, using specialized models tailored to their use case.
Show more
Big day at #GoogleIO#! Bringing together the best of a search engine + the best of AI to help you ask what's really on your mind: + New intelligent Search box + seamless AI Search + Information agents + Antigravity in Search + Universal Cart My run down on what we shared below 👇 💬 New intelligent Search box & seamless Search experience: This is our biggest upgrade to our iconic Search box since its debut 25 years ago! Now, it expands as you type, you can ask questions across modalities (with text, images, files, Chrome tabs, videos), and it offers AI-powered suggestions to help you formulate your *whole* question. We’re also making it simpler to continue the conversation with Search, bringing AI Overviews and AI Mode into a seamless AI Search experience. So you can flow from your question, to your response on the main Search Results Page, to follow-ups in AI Mode. 👀 Information agents to keep you updated on whatever you’re looking for: We’re entering the era of Search agents! To start, you can set information agents to work for you in the background, searching 24/7 to find you exactly what you’re looking for so you can get updated the moment it changes. Your information agents give you intelligent, synthesized updates and hyper-relevant links, so you can discover fresh content from websites and creators and stay on top of whatever matters to you most. 🔍 Bringing Antigravity into Search: We’re bringing Gemini 3.5 Flash’s agentic coding capabilities and Antigravity right into Search. So Search can now custom-code responses, on the fly, totally personalized to your question. You can get interactive visuals to help you understand a tough concept. And you can even ask Search to build you dynamic tools, dashboards, or planners for long-running projects you want to keep coming back to. 🛍️ Introducing the Universal Cart: We've been building the foundation for agentic shopping with the world’s most comprehensive catalog of products in our Shopping Graph, a common language for agents and systems with the Universal Commerce Protocol (UCP), and the payments infrastructure to make agentic checkout seamless. Now we’re introducing the Universal Cart, an intelligent shopping cart that enables you to shop across merchants and will be available across Search, Gemini app, YouTube, and Gmail. With over 1B monthly active users on AI Mode, overall Search queries in Q1 reaching an all-time high, and the new frontier capabilities we announced today, it’s clear that AI is fueling Search’s next wave of growth.
Show more
TL;DR Self-evolving agents that write their own questions and answer them can fall into "co-cheating," where the proposer and solver quietly agree on the same mistakes. Splitting source documents to evaluate across folds fixes this and lifts performance by over 8 points. Title: False Frontiers: Diagnosing and Mitigating Co-Cheating in Self-Evolving Search Agents URL: Points 🔁 Proposer and solver share source-derived errors, letting false agreement cycle back as reward — the paper calls this "co-cheating" 📉 Standard Dr. Zero systems show 6.1% and 8.8% false-agreement mass ✂️ CrossFit splits source documents into two folds, scoring each proposer's questions with a solver trained only on the other fold 📊 CrossFit alone cuts false agreement to 3.0%/3.7%; combined with MSV it drops to 2.0%/1.7% 🚀 Average downstream Cover-EM improves by 8.8 and 8.4 points over Dr. Zero 🧩 Multi-hop tasks see the biggest gains, averaging over 10 points 💰 Compute cost rises 1.72-2.7x over baseline, though a half-budget variant still works What stands out: without auditing the evaluator's own training history, apparent progress can be an illusion. #SelfEvolvingAgents# #ReinforcementLearning#
Show more