Close Menu
Ztoog
    What's Hot
    Technology

    Everything you need to know about the AI chatbot

    AI

    AI-readiness for C-suite leaders | MIT Technology Review

    Science

    Strange quantum effect observed in unusually large object

    Important Pages:
    • About Us
    • Contact us
    • Privacy Policy
    • Terms & Conditions
    Facebook X (Twitter) Instagram Pinterest
    Facebook X (Twitter) Instagram Pinterest
    Ztoog
    • Home
    • The Future

      OPPO launches A5 Pro 5G: Premium features at a budget price

      How I Turn Unstructured PDFs into Revenue-Ready Spreadsheets

      Is it the best tool for 2025?

      The clocks that helped define time from London’s Royal Observatory

      Summer Movies Are Here, and So Are the New Popcorn Buckets

    • Technology

      What It Is and Why It Matters—Part 1 – O’Reilly

      Ensure Hard Work Is Recognized With These 3 Steps

      Cicada map 2025: Where will Brood XIV cicadas emerge this spring?

      Is Duolingo the face of an AI jobs crisis?

      The US DOD transfers its AI-based Open Price Exploration for National Security program to nonprofit Critical Minerals Forum to boost Western supply deals (Ernest Scheyder/Reuters)

    • Gadgets

      Maono Caster G1 Neo & PD200X Review: Budget Streaming Gear for Aspiring Creators

      Apple plans to split iPhone 18 launch into two phases in 2026

      Upgrade your desk to Starfleet status with this $95 USB-C hub

      37 Best Graduation Gift Ideas (2025): For College Grads

      Backblaze responds to claims of “sham accounting,” customer backups at risk

    • Mobile

      Motorola’s Moto Watch needs to start living up to the brand name

      Samsung Galaxy S25 Edge promo materials leak

      What are people doing with those free T-Mobile lines? Way more than you’d expect

      Samsung doesn’t want budget Galaxy phones to use exclusive AI features

      COROS’s charging adapter is a neat solution to the smartwatch charging cable problem

    • Science

      Nothing is stronger than quantum connections – and now we know why

      Failed Soviet probe will soon crash to Earth – and we don’t know where

      Trump administration cuts off all future federal funding to Harvard

      Does kissing spread gluten? New research offers a clue.

      Why Balcony Solar Panels Haven’t Taken Off in the US

    • AI

      Hybrid AI model crafts smooth, high-quality videos in seconds | Ztoog

      How to build a better AI benchmark

      Q&A: A roadmap for revolutionizing health care through data-driven innovation | Ztoog

      This data set helps researchers spot harmful stereotypes in LLMs

      Making AI models more trustworthy for high-stakes settings | Ztoog

    • Crypto

      Ethereum Breaks Key Resistance In One Massive Move – Higher High Confirms Momentum

      ‘The Big Short’ Coming For Bitcoin? Why BTC Will Clear $110,000

      Bitcoin Holds Above $95K Despite Weak Blockchain Activity — Analytics Firm Explains Why

      eToro eyes US IPO launch as early as next week amid easing concerns over Trump’s tariffs

      Cardano ‘Looks Dope,’ Analyst Predicts Big Move Soon

    Ztoog
    Home » Overcoming Hallucinations in AI: How Factually Augmented RLHF Optimizes Vision-Language Alignment in Large Multimodal Models
    AI

    Overcoming Hallucinations in AI: How Factually Augmented RLHF Optimizes Vision-Language Alignment in Large Multimodal Models

    Facebook Twitter Pinterest WhatsApp
    Overcoming Hallucinations in AI: How Factually Augmented RLHF Optimizes Vision-Language Alignment in Large Multimodal Models
    Share
    Facebook Twitter LinkedIn Pinterest WhatsApp

    By extra pre-training utilizing image-text pairings or fine-tuning them with specialised visible instruction tuning datasets, Large Language Models could dive into the multimodal area, giving rise to potent Large Multimodal Models. However, there are obstacles to constructing LMMs, chief amongst them the disparity between the amount and high quality of multimodal information and text-only datasets. Take the LLaVA mannequin, initialized from a pre-trained visible encoder and a language mannequin tweaked for directions. It is educated on far fewer situations than text-only fashions, which use over 100M examples over 1800 duties. It is simply educated on 150K synthetic image-based conversations. Due to such information restrictions, the visible and language modalities will not be aligned. 

    As a consequence, LMMs may generate hallucinatory outputs which are inaccurately tied to the context that footage give. Researchers from UC Berkeley, CMU, UIUC, UW–Madison, UMass Amherst Microsoft Research, and MIT-IBM Watson AI Lab current LLaVA-RLHF, a vision-language mannequin educated for enhanced multimodal alignment, to handle the problems introduced on by the absence of high-quality visible instruction tuning information for LMM coaching. One of their main contributions is adapting the multimodal alignment for LMMs to the common and scalable alignment paradigm referred to as Reinforcement Learning from Human Feedback, which has demonstrated outstanding effectiveness for text-based AI brokers. To fine-tune LMM, it collects human preferences specializing in recognizing hallucinations and makes use of these preferences in reinforcement studying. 

    This technique could enhance the multimodal alignment at a comparatively low cost annotation price, similar to $3000 for gathering 10K human preferences for image-based discussions. As far as they know, this technique is the primary efficient use of RLHF for multimodal alignment. Gaining excessive rankings from the reward mannequin solely generally equates to bettering human judgments, which is reward hacking. It is a potential drawback with the current RLHF paradigm. Previous analysis advised iteratively gathering “fresh” human suggestions to cease incentive hacking, however this methodology is usually costly and can’t correctly use present human desire information. This research suggests a extra data-efficient possibility, trying to make the reward mannequin able to utilizing the information and information already current in greater language fashions that people have annotated. 

    Figure 1: A diagram illustrating the opportunity of hallucinations in the course of the Supervised Fine-Tuning (SFT) part of LMM coaching and the way in which Factually Augmented RLHF addresses the issue of low capability in the reward mannequin, which is initialized from the SFT mannequin.

    First, they use a superior visible encoder with larger resolutions and a much bigger language mannequin to reinforce the reward mannequin’s general performance. Second, they current the Factually Augmented RLHF algorithm, which, as proven in Fig. 1, calibrates the reward indicators by supplementing them with additional data like image descriptions or a ground-truth multi-choice possibility. They additional increase the artificial imaginative and prescient instruction tuning information with present high-quality human-annotated multimodal information in the dialog format to reinforce the overall capabilities of LMMs in the course of the Supervised Fine-Tuning stage. They particularly remodel Flickr30k right into a Spotting Captioning project, VQA-v2, and A-OKVQA right into a multi-round QA process, and each prepare the LLaVA-SFT+ fashions utilizing the brand new information set. 

    Finally, they contemplate the right way to consider the multimodal alignment of LMMs in conditions of real-world creation, paying explicit consideration to penalizing any hallucinations. The benchmark questions they develop, MMHAL-BENCH, cowl all 12 of COCO’s key object classes and comprise eight job sorts. According to their evaluation, this benchmark dataset intently matches human assessments, particularly if scores are thought of for anti-hallucinations. As the primary LMM educated with RLHF, LLaVA-RLHF performs admirably in their experimental evaluation. They noticed an enchancment of 94% on the LLaVA-Bench, a 60% enchancment on the MMHAL-BENCH, and so they set new efficiency information for LLaVA with 52.4% on MMBench and 82.7% F1 on POPE. On GitHub, they’ve made their code, mannequin, and information accessible to the general public.


    Check out the Paper and Project. All Credit For This Research Goes To the Researchers on This Project. Also, don’t overlook to affix our 31k+ ML SubReddit, 40k+ Facebook Community, Discord Channel, and Email Newsletter, the place we share the newest AI analysis information, cool AI tasks, and extra.

    If you want our work, you’ll love our e-newsletter..


    Aneesh Tickoo is a consulting intern at MarktechPost. He is at the moment pursuing his undergraduate diploma in Data Science and Artificial Intelligence from the Indian Institute of Technology(IIT), Bhilai. He spends most of his time engaged on tasks aimed toward harnessing the ability of machine studying. His analysis curiosity is picture processing and is captivated with constructing options round it. He loves to attach with folks and collaborate on attention-grabbing tasks.


    ▶️ Now Watch AI Research Updates On Our Youtube Channel [Watch Now]

    Share. Facebook Twitter Pinterest LinkedIn WhatsApp

    Related Posts

    AI

    Hybrid AI model crafts smooth, high-quality videos in seconds | Ztoog

    AI

    How to build a better AI benchmark

    AI

    Q&A: A roadmap for revolutionizing health care through data-driven innovation | Ztoog

    AI

    This data set helps researchers spot harmful stereotypes in LLMs

    AI

    Making AI models more trustworthy for high-stakes settings | Ztoog

    AI

    The AI Hype Index: AI agent cyberattacks, racing robots, and musical models

    AI

    Novel method detects microbial contamination in cell cultures | Ztoog

    AI

    Seeing AI as a collaborator, not a creator

    Leave A Reply Cancel Reply

    Follow Us
    • Facebook
    • Twitter
    • Pinterest
    • Instagram
    Top Posts
    Mobile

    Google Play Store to let you uninstall apps remotely from other devices

    Play Store model 38.8 is out and Google is including the flexibility to uninstall apps…

    The Future

    Disney’s Truly Wild 100th Anniversary Year

    Image: Walt Disney PicturesThe House of Mouse celebrated its centennial in 2023 via monumental movie…

    Science

    Strange alien worlds suggest Earth could survive the death of the sun

    MARK GARLICK/SCIENCE PHOTO LIBRARY IT ALL appeared so easy after we knew the date of…

    Crypto

    Robinhood is on a quest to dive deeper into crypto

    It desires to appeal to the “entire spectrum” of digital asset traders Jacquelyn Melinek 21…

    The Future

    BlueAnt’s new X3i refines an already impressive option

    Some critiques come and go, with out an excessive amount of by the use of…

    Our Picks
    Science

    Remains of planet that formed the Moon may be hiding near Earth’s core

    Technology

    Chinese Malware Hits Systems on Guam. Is Taiwan the Real Target?

    Crypto

    Bridging the Gap Between Solana and Ethereum: Neon EVM Debuts on Solana Mainnet

    Categories
    • AI (1,483)
    • Crypto (1,745)
    • Gadgets (1,796)
    • Mobile (1,840)
    • Science (1,854)
    • Technology (1,790)
    • The Future (1,636)
    Most Popular
    Mobile

    OnePlus Nord N30 5G runs Geekbench with Snapdragon 695 SoC on board

    Mobile

    Black Friday deals: Amazon UK offering deep discounts on Pixel 7 series, Galaxy Z Fold4 and more

    Mobile

    Meta Quest 3’s more affordable sibling may have leaked in new render

    Ztoog
    Facebook X (Twitter) Instagram Pinterest
    • Home
    • About Us
    • Contact us
    • Privacy Policy
    • Terms & Conditions
    © 2025 Ztoog.

    Type above and press Enter to search. Press Esc to cancel.