index.html

<!doctype html>
<html lang="en">
<head>
    <meta charset="utf-8">
    <meta name="viewport" content="width=device-width, initial-scale=1">
    <meta name="description" content="Ninghui Feng is an AI researcher and engineer working on trustworthy AI, multimodal reasoning, time-series intelligence, and generative media systems.">
    <meta name="author" content="Ninghui Feng">
    <meta name="theme-color" content="#ffffff">

    <title>Ninghui Feng's Homepage</title>
    <link rel="canonical" href="https://feng1201.github.io/">
    <link rel="icon" type="image/x-icon" href="static/assets/favicon.ico">

    <meta property="og:type" content="website">
    <meta property="og:url" content="https://feng1201.github.io/">
    <meta property="og:title" content="Ninghui Feng's Homepage">
    <meta property="og:description" content="Research on trustworthy AI, multimodal reasoning, time-series intelligence, and generative media systems.">
    <meta property="og:image" content="https://feng1201.github.io/static/assets/img/og-minimal.png">
    <meta name="twitter:card" content="summary_large_image">
    <meta name="twitter:title" content="Ninghui Feng's Homepage">
    <meta name="twitter:description" content="Research on trustworthy AI, multimodal reasoning, time-series intelligence, and generative media systems.">
    <meta name="twitter:image" content="https://feng1201.github.io/static/assets/img/og-minimal.png">

    <link rel="stylesheet" href="static/css/main.css">
    <script src="static/js/scripts.js" defer></script>

    <script type="application/ld+json">
    {
        "@context": "https://schema.org",
        "@type": "Person",
        "name": "Ninghui Feng",
        "url": "https://feng1201.github.io/",
        "email": "mailto:feng65501@gmail.com",
        "jobTitle": "AI Researcher and Engineer",
        "affiliation": {
            "@type": "CollegeOrUniversity",
            "name": "University of Nottingham Ningbo China"
        },
        "alumniOf": {
            "@type": "CollegeOrUniversity",
            "name": "Northeast Electric Power University"
        },
        "sameAs": [
            "https://scholar.google.com/citations?user=YpJ-U1sAAAAJ&hl=en"
        ]
    }
    </script>
</head>

<body>
    <a class="skip-link" href="#main">Skip to content</a>

    <header class="site-header">
        <div class="masthead">
            <a class="site-brand" href="#top" aria-label="Ninghui Feng's homepage">
                <span class="brand-mark" aria-hidden="true">NF</span>
                <span>Ninghui Feng's Homepage</span>
            </a>

            <button class="menu-button" type="button" aria-expanded="false" aria-controls="site-nav" data-menu-button>
                <span class="menu-label">Menu</span>
                <span class="menu-icon" aria-hidden="true"></span>
            </button>

            <nav class="site-nav" id="site-nav" aria-label="Primary navigation" data-menu>
                <a href="#publications">Publications</a>
                <a href="#experience">Experience</a>
                <a href="#education">Education</a>
                <a href="#snapshots">Snapshots of Life</a>
                <a href="/blog/">Blog</a>
            </nav>
        </div>
    </header>

    <main class="page-layout" id="main">
        <aside class="profile" id="top">
            <div class="profile-heading">
                <figure class="profile-photo">
                    <img class="avatar" src="static/assets/img/avatar-haikou.jpg" alt="Portrait of Ninghui Feng in Haikou" width="525" height="720">
                    <figcaption>Taken in Haikou on New Year's Eve, welcoming 2026.</figcaption>
                </figure>
                <div>
                    <h1>Ninghui Feng</h1>
                    <p>MRes Student, University of Nottingham Ningbo China<br>Agent Application R&amp;D Intern, LuckyShort</p>
                </div>
            </div>

            <ul class="profile-links" aria-label="Contact and academic profiles">
                <li>
                    <a href="mailto:feng65501@gmail.com">
                        <span aria-hidden="true">@</span> Email
                    </a>
                </li>
                <li>
                    <a href="https://scholar.google.com/citations?user=YpJ-U1sAAAAJ&hl=en" target="_blank" rel="noopener noreferrer">
                        <span aria-hidden="true">S</span> Google Scholar
                    </a>
                </li>
                <li>
                    <a href="mailto:Ninghui.FENG@nottingham.edu.cn">
                        <span aria-hidden="true">U</span> University Email
                    </a>
                </li>
            </ul>
        </aside>

        <article class="content">
            <section class="content-section biography" aria-labelledby="biography-title">
                <h2 id="biography-title">Biography</h2>

                <p>
                    I am an MRes student in Electrical Engineering at the
                    <a href="https://www.nottingham.edu.cn/" target="_blank" rel="noopener noreferrer">University of Nottingham Ningbo China</a>.
                    I am also an Agent Application R&amp;D Intern at <strong>LuckyShort</strong>, where I work on video-algorithm applications and agent workflows for AI-native content production.
                </p>

                <p>
                    I work on <strong>trustworthy artificial intelligence</strong>. My research interests broadly lie in multimodal large language models, numerical and time-series reasoning, reliable forecasting, and generative media systems. I am particularly interested in models that can recognize uncertainty, explain their reasoning, and remain dependable when real-world data is noisy or unfamiliar.
                </p>

                <p>
                    Previously, I was an Algorithm R&amp;D Intern at the
                    <a href="https://www.idea.edu.cn/" target="_blank" rel="noopener noreferrer">International Digital Economy Academy (IDEA)</a>
                    and a Research Assistant at the
                    <a href="https://www.hkust-gz.edu.cn/" target="_blank" rel="noopener noreferrer">Hong Kong University of Science and Technology (Guangzhou)</a>.
                    My work has appeared at venues including EMNLP Findings, ICML, ACM Multimedia, ICRA, ICASSP, IEEE Sensors Journal, and Pattern Recognition.
                </p>

                <p class="note">
                    I am open to research collaborations and engineering discussions on trustworthy AI, multimodal reasoning, time-series intelligence, and agentic content systems.
                </p>
            </section>

            <section class="content-section" aria-labelledby="news-title">
                <h2 id="news-title">Recent News</h2>

                <ul class="news-list">
                    <li><time datetime="2026-09">September 2026:</time> Our <a href="https://github.com/langroid/langroid" target="_blank" rel="noopener noreferrer">Langroid</a> has surpassed 4,000 GitHub stars!</li>
                    <li><time datetime="2026-08">August 2026:</time> Our paper <em>NUM2EVENT</em> was accepted to EMNLP 2026 Findings.</li>
                    <li><time datetime="2026-03">March 2026:</time> Joined LuckyShort as an Agent Application R&amp;D Intern.</li>
                    <li><time datetime="2026">2026:</time> Our work on reliable time-series forecasting under future uncertainty was accepted to ICASSP 2026.</li>
                    <li><time datetime="2025-10">October 2025:</time> Our paper on faithful multimedia web forecasting was published at ACM Multimedia 2025.</li>
                    <li><time datetime="2025-08">August 2025:</time> Joined IDEA as an Algorithm R&amp;D Intern, focusing on multimodal LLM post-training.</li>
                    <li><time datetime="2025">2025:</time> Co-authored papers were published at ICML 2025 and ICRA 2025.</li>
                    <li><time datetime="2024-08">August 2024:</time> Joined HKUST(GZ) as a Research Assistant.</li>
                </ul>
            </section>

            <section class="content-section" id="publications" aria-labelledby="publications-title">
                <div class="section-heading-row">
                    <h2 id="publications-title">Selected Publications</h2>
                    <a href="https://scholar.google.com/citations?user=YpJ-U1sAAAAJ&hl=en" target="_blank" rel="noopener noreferrer">Full list on Google Scholar</a>
                </div>

                <div class="publication-list">
                    <article class="publication">
                        <p class="paper-title">
                            <span>[EMNLP Findings '26]</span>
                            <a href="https://openreview.net/forum?id=6uM9uNwekL" target="_blank" rel="noopener noreferrer"><em>NUM2EVENT: Structured Event Inference from Numerical Time-Series</em></a>
                            <a class="paper-link" href="https://openreview.net/forum?id=6uM9uNwekL" target="_blank" rel="noopener noreferrer">Paper</a>
                        </p>
                        <p class="authors"><strong>Ninghui Feng</strong>, Songning Lai, Xinan Chen, Zheng Chu, Yiyan Qi</p>
                        <p class="highlight"><strong>Highlight:</strong> A number-to-event framework that infers structured, interpretable events directly from numerical signals.</p>
                    </article>

                    <article class="publication">
                        <p class="paper-title">
                            <span>[ICASSP '26]</span>
                            <a href="https://arxiv.org/abs/2503.19656" target="_blank" rel="noopener noreferrer"><em>Towards Reliable Time Series Forecasting under Future Uncertainty: Ambiguity and Novelty Rejection Mechanisms</em></a>
                            <a class="paper-link" href="https://arxiv.org/abs/2503.19656" target="_blank" rel="noopener noreferrer">Paper</a>
                        </p>
                        <p class="authors"><strong>Ninghui Feng</strong>, Songning Lai, Xin Zhou, Jiayu Yang, et al.</p>
                        <p class="highlight"><strong>Highlight:</strong> Reliable forecasting that can identify ambiguous predictions and reject unfamiliar future conditions.</p>
                    </article>

                    <article class="publication">
                        <p class="paper-title">
                            <span>[ACM MM '25]</span>
                            <a href="https://xll0328.github.io/fts/" target="_blank" rel="noopener noreferrer"><em>From Guesswork to Guarantee: Towards Faithful Multimedia Web Forecasting with TimeSieve</em></a>
                            <a class="paper-link" href="https://xll0328.github.io/fts/" target="_blank" rel="noopener noreferrer">Project</a>
                        </p>
                        <p class="authors">Songning Lai*, <strong>Ninghui Feng*</strong>, Jiechao Gao, Hao Wang, et al.</p>
                        <p class="highlight"><strong>Highlight:</strong> Faithfulness-aware optimization that improves robustness to random seeds and several kinds of model perturbation.</p>
                    </article>

                    <article class="publication">
                        <p class="paper-title">
                            <span>[ICML '25]</span>
                            <a href="https://proceedings.mlr.press/v267/hu25y.html" target="_blank" rel="noopener noreferrer"><em>IMTS is Worth Time × Channel Patches: Visual Masked Autoencoders for Irregular Multivariate Time Series Prediction</em></a>
                            <a class="paper-link" href="https://proceedings.mlr.press/v267/hu25y.html" target="_blank" rel="noopener noreferrer">Paper</a>
                        </p>
                        <p class="authors">Zhangyi Hu*, Jiemin Wu*, Hua Xu*, Mingqian Liao, <strong>Ninghui Feng</strong>, Bo Gao, Songning Lai, Yutao Yue</p>
                        <p class="highlight"><strong>Highlight:</strong> Visual masked autoencoders that model irregular multivariate time series through time-by-channel patches.</p>
                    </article>

                    <article class="publication">
                        <p class="paper-title">
                            <span>[IEEE Sensors Journal]</span>
                            <a href="https://scholar.google.com/citations?view_op=view_citation&hl=en&user=YpJ-U1sAAAAJ&citation_for_view=YpJ-U1sAAAAJ:UeHWp8X0CEIC" target="_blank" rel="noopener noreferrer"><em>Determination of Tea Quality: A Separate Domain Graph Convolution Network Combined With an Electronic Nose</em></a>
                            <a class="paper-link" href="https://scholar.google.com/citations?view_op=view_citation&hl=en&user=YpJ-U1sAAAAJ&citation_for_view=YpJ-U1sAAAAJ:UeHWp8X0CEIC" target="_blank" rel="noopener noreferrer">Paper</a>
                        </p>
                        <p class="authors"><strong>Ninghui Feng</strong>, Siyuan Kang, Chongbo Yin, Ziyang Li, Yan Shi</p>
                        <p class="highlight"><strong>Highlight:</strong> Graph-based feature learning for tea-quality recognition from electronic-nose sensor data.</p>
                    </article>

                    <article class="publication">
                        <p class="paper-title">
                            <span>[Preprint '24]</span>
                            <a href="https://arxiv.org/abs/2406.05036" target="_blank" rel="noopener noreferrer"><em>TimeSieve: Extracting Temporal Dynamics through Information Bottlenecks</em></a>
                            <a class="paper-link" href="https://arxiv.org/abs/2406.05036" target="_blank" rel="noopener noreferrer">Paper</a>
                        </p>
                        <p class="authors"><strong>Ninghui Feng</strong>, Songning Lai, Jiayu Yang, Fobao Zhou, Zhenxiao Yin, Hang Zhao</p>
                        <p class="highlight"><strong>Highlight:</strong> Wavelet decomposition and information bottlenecks for learning useful multi-scale temporal patterns.</p>
                    </article>
                </div>

                <p class="author-note">* indicates equal contribution.</p>
            </section>

            <section class="content-section" id="experience" aria-labelledby="experience-title">
                <h2 id="experience-title">Experience</h2>

                <div class="simple-list">
                    <article>
                        <time>Mar. 2026 — Present</time>
                        <div>
                            <h3>Agent Application R&amp;D Intern, LuckyShort</h3>
                            <p>Video generation, agent workflows, production monitoring, and AI short-drama tooling.</p>
                        </div>
                    </article>
                    <article>
                        <time>Aug. 2025 — Feb. 2026</time>
                        <div>
                            <h3>Algorithm R&amp;D Intern, IDEA</h3>
                            <p>Multimodal LLM post-training and structured reasoning from numerical time series.</p>
                        </div>
                    </article>
                    <article>
                        <time>Aug. 2024 — Jul. 2025</time>
                        <div>
                            <h3>Research Assistant, HKUST(GZ)</h3>
                            <p>Trustworthy AI, time-series forecasting, and non-intrusive load monitoring.</p>
                        </div>
                    </article>
                    <article>
                        <time>Jul. 2021 — Jan. 2024</time>
                        <div>
                            <h3>Undergraduate Researcher, Bionic Sensing &amp; Pattern Recognition Research Team</h3>
                            <p>Graph neural networks and machine olfaction.</p>
                        </div>
                    </article>
                </div>

                <details class="project-details">
                    <summary>Selected engineering impact at LuckyShort</summary>
                    <p>
                        I contributed to script parsing, character assets, video generation, super-resolution, background music, monitoring, and delivery tooling. The median script-processing time decreased from <strong>232 seconds to 145 seconds</strong>: a typical task became 87 seconds faster, which is a 37.5% reduction. The BGM failure rate decreased by <strong>35% relative to the previous pipeline</strong>, and the rollout data pipeline captured <strong>37,000</strong> real generate and regenerate interactions for analysis.
                    </p>
                </details>
            </section>

            <section class="content-section" id="education" aria-labelledby="education-title">
                <h2 id="education-title">Education &amp; Awards</h2>

                <div class="simple-list education-list">
                    <article>
                        <time>2025 — 2026</time>
                        <div>
                            <h3>MRes in Electrical Engineering</h3>
                            <p>University of Nottingham Ningbo China</p>
                        </div>
                    </article>
                    <article>
                        <time>2020 — 2024</time>
                        <div>
                            <h3>B.Eng. in Measurement &amp; Control Technology and Instruments</h3>
                            <p>Northeast Electric Power University</p>
                        </div>
                    </article>
                </div>

                <ul class="awards-list">
                    <li>National College Students' Innovation and Entrepreneurship Training Program project for a real-time meat-quality analysis system based on machine olfaction.</li>
                    <li>China Southern Power Grid 2024 Smart Metering Ecosystem Development Competition.</li>
                </ul>
            </section>

            <section class="content-section" id="snapshots" aria-labelledby="snapshots-title">
                <h2 id="snapshots-title">Snapshots of Life</h2>

                <p>
                    Outside work, skiing is one of my favorite ways to spend winter. During my undergraduate years at Northeast Electric Power University, Tianding Mountain Ski Resort was only about an hour away.
                </p>

                <p>
                    At the beginning of 2026, I travelled with classmates to Haikou and Sanya, Hainan. The trip gave us time to enjoy the coast and start the year together.
                </p>

                <div class="travel-gallery" aria-label="Photos from Hainan">
                    <figure>
                        <img src="static/assets/img/hainan-coast.jpg" alt="Sunlight breaking through clouds above the South China Sea in Sanya" width="1280" height="960" loading="lazy" decoding="async">
                        <figcaption>The South China Sea, Sanya · Early 2026</figcaption>
                    </figure>
                    <figure>
                        <img src="static/assets/img/nanshan-guanyin.jpg" alt="Statue of Guanyin of the South Sea against a blue sky in Sanya" width="963" height="1280" loading="lazy" decoding="async">
                        <figcaption>Guanyin of the South Sea, Sanya · Early 2026</figcaption>
                    </figure>
                </div>
            </section>
        </article>
    </main>

    <footer class="site-footer">
        <div>
            <p>© <span data-year>2026</span> Ninghui Feng.</p>
            <p>
                <a href="mailto:feng65501@gmail.com">Email</a>
                <span>·</span>
                <a href="https://scholar.google.com/citations?user=YpJ-U1sAAAAJ&hl=en" target="_blank" rel="noopener noreferrer">Google Scholar</a>
            </p>
        </div>
    </footer>
</body>
</html>