Hasil untuk "Vocational guidance. Career development"

Kwanyoung Kim

Classifier-Free Guidance (CFG) has significantly enhanced the generative quality of diffusion models by extrapolating between conditional and unconditional outputs. However, its high inference cost and limited applicability to distilled or single-step models have shifted research focus toward attention-space extrapolation. While these methods offer computational efficiency, their theoretical underpinnings remain elusive. In this work, we establish a foundational framework for attention-space extrapolation by modeling attention dynamics as fixed-point iterations within Modern Hopfield Networks. We demonstrate that the extrapolation effect in attention space constitutes a special case of Anderson Acceleration applied to these dynamics. Building on this insight and the weak contraction property, we propose Geometry Aware Attention Guidance (GAG). By decomposing attention updates into parallel and orthogonal components relative to the guidance direction, GAG stabilizes the acceleration process and maximizes guidance efficiency. Our plug-and-play method seamlessly integrates with existing frameworks while significantly improving generation quality.

Aglaia Stampoltzis, Eleni Peristeri, Rany Kalouri

BackgroundIndividuals with Autism Spectrum Disorder (ASD) possess a unique range of strengths and challenges that can impact their employment opportunities and their vocational outcomes. Career counselors' role in helping individuals with ASD to their transition to employment has not been fully understood.ObjectiveThe aim of the current study was threefold: (a) to explore career counselors' views and attitudes toward employability skills in transition aged individuals with ASD in Greece; (b) to investigate the counselors' perception of the challenges they face when working with this population; and (c) to highlight career counselors' judgment of the suitability of professions for autistic individuals. For the first and second aim, we used an exploratory and confirmatory factor analysis.MethodsAn original 28-item survey was developed and disseminated to career counselors. A total of 92 professionals (62 women) took part in the study. All of them have been working as career counselors in the public or private sector. The factor structure of the survey's items was examined using quantitative data analysis, namely, an exploratory and confirmatory factor analyses factor method.ResultsAccording to the results of the exploratory and confirmatory factor analysis, as well as descriptive statistics, we found that counselors agreed that social competence and high self-esteem can promote positive professional development in people with ASD, and that technology can have positive effects in their career. Over half of the counselors surveyed think there are professions particularly well-suited to individuals with ASD and they expressed a strong desire for ASD-specific training to be better prepared to meet the needs of their clients.ConclusionThe results of the study represent the first step toward key variables in vocational guidance for individuals with ASD in Greece that can guide future research.

Boseong Jeon

This paper presents a test-time guidance method to improve the output quality of the human motion diffusion models without requiring additional training. To have negative guidance, Smooth Perturbation Guidance (SPG) builds a weak model by temporally smoothing the motion in the denoising steps. Compared to model-agnostic methods originating from the image generation field, SPG effectively mitigates out-of-distribution issues when perturbing motion diffusion models. In SPG guidance, the nature of motion structure remains intact. This work conducts a comprehensive analysis across distinct model architectures and tasks. Despite its extremely simple implementation and no need for additional training requirements, SPG consistently enhances motion fidelity. Project page can be found at https://spg-blind.vercel.app/

Ky Dan Nguyen, Hoang Lam Tran, Anh-Dung Dinh et al.

Autoregressive (AR) models based on next-scale prediction are rapidly emerging as a powerful tool for image generation, but they face a critical weakness: information inconsistencies between patches across timesteps introduced by progressive resolution scaling. These inconsistencies scatter guidance signals, causing them to drift away from conditioning information and leaving behind ambiguous, unfaithful features. We tackle this challenge with Information-Grounding Guidance (IGG), a novel mechanism that anchors guidance to semantically important regions through attention. By adaptively reinforcing informative patches during sampling, IGG ensures that guidance and content remain tightly aligned. Across both class-conditioned and text-to-image generation tasks, IGG delivers sharper, more coherent, and semantically grounded images, setting a new benchmark for AR-based methods.

Yang Yue, Yulin Wang, Haojun Jiang et al.

Echocardiography is crucial for cardiovascular disease detection but relies heavily on experienced sonographers. Echocardiography probe guidance systems, which provide real-time movement instructions for acquiring standard plane images, offer a promising solution for AI-assisted or fully autonomous scanning. However, developing effective machine learning models for this task remains challenging, as they must grasp heart anatomy and the intricate interplay between probe motion and visual signals. To address this, we present EchoWorld, a motion-aware world modeling framework for probe guidance that encodes anatomical knowledge and motion-induced visual dynamics, while effectively leveraging past visual-motion sequences to enhance guidance precision. EchoWorld employs a pre-training strategy inspired by world modeling principles, where the model predicts masked anatomical regions and simulates the visual outcomes of probe adjustments. Built upon this pre-trained model, we introduce a motion-aware attention mechanism in the fine-tuning stage that effectively integrates historical visual-motion data, enabling precise and adaptive probe guidance. Trained on more than one million ultrasound images from over 200 routine scans, EchoWorld effectively captures key echocardiographic knowledge, as validated by qualitative analysis. Moreover, our method significantly reduces guidance errors compared to existing visual backbones and guidance frameworks, excelling in both single-frame and sequential evaluation protocols. Code is available at https://github.com/LeapLabTHU/EchoWorld.

Chase Hatcher, Adrienne Traxler, Lily Donis et al.

This paper presents a social network analysis of the professional support networks of 100 LGBTQ+ and/or women PhD physicists, comparing the networks based on the career sectors of academia, industry, and government/nonprofit. The methods for constructing and analyzing the ego networks, which are novel in many ways, are explained in greater detail in an earlier publication (Hatcher et. al., 2025). We use statistical tests of independence to explore differences between sectors in terms of whole network metrics, network composition based on alter characteristics, and support types. We find that alters associated with groups (like affinity groups and personal and professional interest groups) are more likely to provide identity-based and community building support, participants in Academia have fewer personal friends in their networks while those in Industry have more, participants in Government report less instrumental support, and those in Academia report less material support. These results and others lead to suggestions for employers in these sectors on how to better support these physicists, including continuing to promote participation in affinity and interest groups, providing more material support and/or personal time in the academic sector, and more instrumental support in the form of professional development or training in the government sector.

Ruitong Sun, Tianze Yang, Wei Niu et al.

Diffusion models have emerged as the dominant paradigm for high-quality image generation, yet their computational expense remains substantial due to iterative denoising. Classifier-Free Guidance (CFG) significantly enhances generation quality and controllability but doubles the computation by requiring both conditional and unconditional forward passes at every timestep. We present OUSAC (Optimized gUidance Scheduling with Adaptive Caching), a framework that accelerates diffusion transformers (DiT) through systematic optimization. Our key insight is that variable guidance scales enable sparse computation: adjusting scales at certain timesteps can compensate for skipping CFG at others, enabling both fewer total sampling steps and fewer CFG steps while maintaining quality. However, variable guidance patterns introduce denoising deviations that undermine standard caching methods, which assume constant CFG scales across steps. Moreover, different transformer blocks are affected at different levels under dynamic conditions. This paper develops a two-stage approach leveraging these insights. Stage-1 employs evolutionary algorithms to jointly optimize which timesteps to skip and what guidance scale to use, eliminating up to 82% of unconditional passes. Stage-2 introduces adaptive rank allocation that tailors calibration efforts per transformer block, maintaining caching effectiveness under variable guidance. Experiments demonstrate that OUSAC significantly outperforms state-of-the-art acceleration methods, achieving 53% computational savings with 15% quality improvement on DiT-XL/2 (ImageNet 512x512), 60% savings with 16.1% improvement on PixArt-alpha (MSCOCO), and 5x speedup on FLUX while improving CLIP Score over the 50-step baseline.

عابدین دارابی عمارتی, زهرا رحمانی صدیق

هدف: هدف این پژوهش، تعیین میزان برازش مدل مفهومی درگیری شغلی معلمان بر اساس عدم تعادل تلاش- پاداش با میانجیگری سلامت معنوی با مدل تجربی است. روش: روش پژوهش توصیفی از نوع همبستگی بود. جامعه آماری شامل کلیه معلمان زن و مرد مدارس دخترانه متوسطه دوم ناحیهی یک شهرستان کرمانشاه به تعداد 313 نفر که از این تعداد 265 نفر زن و 48 نفر مرد بود. به روش نمونهگیری تصادفی از نوع طبقهای تعداد 142 نفر زن و 27 نفر مرد بهعنوان نمونه آماری انتخاب شد. برای گردآوری دادهها از سه پرسشنامه عدم تعادل تلاش- پاداش (ERIQ) سیگریست (2013)، سلامت معنوی سایه میری و همکاران (1395) و درگیری شغلی کانونگو (1982) استفاده شد. بهمنظور تحلیل دادهها، از مدلیابی معادلات ساختاری با استفاده از نرمافزار SPSS و AMOS استفاده شد. یافتهها: نتایج پژوهش نشان داد که عدم تعادل تلاش-پاداش بر سلامت معنوی و درگیری شغلی اثر معناداری دارد. سلامت معنوی بر درگیری شغلی اثر معناداری دارد. همچنین متغیر سلامت معنوی نقش میانجی بر اثر عدم تعادل تلاش-پاداش بر درگیری شغلی را دارد. همچنین شاخصهای برازندگی مدل حاکی از آن بود که مدل ارائهشده از برازش مناسبی برخوردار است. نتیجهگیری: با توجه به وجود اثر عدم تعادل تلاش-پاداش و سلامت معنوی معلمان بر درگیری شغلی، میتوان نتیجه گرفت که با تعادل تلاش-پاداش و ارتقای سطح سلامت معنوی معلمان، درگیری شغلی بهبود مییابد.

کبری طرفه ئی منفرد, مرجان واحدی, رویا اشراقی سامانی et al.

در سال‌های اخیر، روستاها به نقاط استراتژیکی برای اجرای طرح‌های مربوط به انرژی تجدیدپذیر تبدیل شده‌اند؛ چرا که این مناطق از پتانسیل بالایی برای برای استقرار تأسیسات مربوط به انرژی تجدیدپذیر و در نتیجه توسعه پایدار روستایی برخوردار هستند. علی‌رغم آن‌که استراتژی‌های سیاستی اثرات مثبت انرژی‌های تجدیدپذیر در دستیابی به توسعه پایدار روستایی را پیش‌بینی نموده‌اند، چگونگی دستیابی به این مهم در تطابق با شرایط فعلی جوامع مختلف، هم‌چنان در هاله‌ای از ابهام به سر می‌برد. در این راستا، پژوهش حاضر با هدف شناسایی عوامل مؤثر بر پیاده سازی و توسعه کاربرد انرژی‌های تجدیدپذیر در مناطق روستایی استان ایلام انجام شد. برای گردآوری داده‌ها در این پژوهش کمی از پرسشنامۀ محقق‌ساخته‌ای استفاده شده که روایی و پایایی آن با استفاده از روایی محتوا و آزمون آلفای کرونباخ مورد تأیید قرار گرفت. جامعه آماری این پژوهش، شامل کلیۀ متخصصان و صاحبنظران حوزۀ انرژی در استان ایلام به تعداد 265 نفر بودند که با استفاده از جدول نمونه‌گیری بارتلت و همکاران، تعداد 154 نفر از آنان به روش تصادفی به عنوان نمونه‌های پژوهش انتخاب شدند. تجزیه و تحلیل داده‌ها با استفاده از نرم‌افزارSPSS نشان داد عوامل حاکمیتی- سیاستی، اقتصادی- مالی، مدیریتی- سازمانی، اجتماعی- فرهنگی، آموزشی- ترویجی، بین‌المللی، ویژگی‌های فنی، کالبدی- زیرساختی، طبیعی- منطقه‌ای و شرایط بازار در سطحی بالاتر از متوسط در پیاده‌سازی و توسعه کاربرد انرژی‌های تجدیدپذیر در مناطق روستایی استان ایلام نقش دارند. نتایج این مطالعه می‌تواند نقش مؤثری در تغییر الگوی تولید انرژی در استان ایلام و حرکت به سمت توسعه انرژی پایدار داشته باشد.

عاطفه اسفندیاری مقدم, محمد ملک زاده, شیرعلی خرامین

هدف پژوهش حاضر با هدف مدل‌یابی نقش میانجی حمایت اجتماعی ادراک شده در رابطه بین تعارض کار-خانواده با فرسودگی شغلی در کارکنان اقماری شرکت راه‌اندازی و بهره‌برداری صنایع نفت صورت پذیرفت. پژوهش حاضر جز تحقیقات توصیفی و از نوع معادلات ساختاری بود. جامعه آماری کارکنان اقماری شرکت راه‌اندازی و بهره-برداری صنایع نفت(اُیکو) در سه ماهه بهار سال 1402 به تعداد 4500 نفر بودند. حجم نمونه بر اساس فرمول کوکران و با روش نمونه‏گیری در دسترس 500 نفر انتخاب شدند. برای جمع‌آوری داده‌ها از سیاهه فرسودگی شغلی مسلاچ(MBI) مسلاچ، جکسون، لیتر، اسچافلی و اسچواب(1986)، پرسشنامه تعارض کار-خانواده(WFCQ) کارلسون، کاکمار و ویلیامز (2000) و مقیاس چندبعدی حمایت اجتماعی ادراک شده(MSPSS) زیمت، داهلم، زیمت و فارلی(1988) استفاده شد. داده‌ها با استفاده از معادلات ساختاری با نرم‌افزار SPSS و AMOS نسخه 28 تجزیه و تحلیل شدند. یافته‌های این پژوهش حاکی از آن بود تعارض کار-خانواده (59/0=β و 001/0=sig) بر فرسودگی شغلی اثر مستقیم و معنادار دارد. در نهایت حمایت اجتماعی ادراک شده(19/0- =β و 05/0>p) دارای اثر مستقیم و معنادار بر فرسودگی شغلی بود و توانست در رابطه بین تعارض کار-خانواده با فرسودگی شغلی نقش میانجی معنادار ایفا کند. همچنین مدل نهایی پژوهش از برازش مطلوبی برخوردار بود (018/0=RMSEA و 05/0>p) و 52 درصد فرسودگی شغلی ‏تبیین می‏شود. می‌توان نتیجه‏گیری کرد که با توجه به نقش میانجی معنادار حمایت اجتماعی ادراک شده، می‏توان با به کارگیری مداخلات خانواده محور از بروز فرسودگی شغلی در کارکنان اقماری پیشگیری کرد.

Lluis Danus, Robert H. Davis, Roger Guimera et al.

We study the influence that research environments have in shaping careers of early-career faculty in terms of their research portfolio. We find that departments exert an attractive force over early-career newcomer faculty, who after their incorporation increase their within-department collaborations, and work on topics closer to those of incumbent faculty. However, these collaborations are not gender blind: Newcomers collaborate less than expected with female senior incumbents. The analysis of departments grouped by fraction of female incumbents reveals that female newcomers in departments with above the median fractions of female incumbents tend to select research topics farther from their department than female newcomers in the remaining departments -- a difference we do not observe for male newcomers. Our results suggest a relationship between the collaboration deficit with female incumbents and the selection of research topics of female early-faculty, thus highlighting the importance of studying research environments to fully understand gender differences in academia.

Geonsun Lee, Dae Yeol Lee, Guan-Ming Su et al.

In this paper, we present a novel multi-modal attention guidance method designed to address the challenges of turn-taking dynamics in meetings and enhance group conversations within virtual reality (VR) environments. Recognizing the difficulties posed by a confined field of view and the absence of detailed gesture tracking in VR, our proposed method aims to mitigate the challenges of noticing new speakers attempting to join the conversation. This approach tailors attention guidance, providing a nuanced experience for highly engaged participants while offering subtler cues for those less engaged, thereby enriching the overall meeting dynamics. Through group interview studies, we gathered insights to guide our design, resulting in a prototype that employs "light" as a diegetic guidance mechanism, complemented by spatial audio. The combination creates an intuitive and immersive meeting environment, effectively directing users' attention to new speakers. An evaluation study, comparing our method to state-of-the-art attention guidance approaches, demonstrated significantly faster response times (p < 0.001), heightened perceived conversation satisfaction (p < 0.001), and preference (p < 0.001) for our method. Our findings contribute to the understanding of design implications for VR social attention guidance, opening avenues for future research and development.

Hyungjin Chung, Jeongsol Kim, Geon Yeong Park et al.

Classifier-free guidance (CFG) is a fundamental tool in modern diffusion models for text-guided generation. Although effective, CFG has notable drawbacks. For instance, DDIM with CFG lacks invertibility, complicating image editing; furthermore, high guidance scales, essential for high-quality outputs, frequently result in issues like mode collapse. Contrary to the widespread belief that these are inherent limitations of diffusion models, this paper reveals that the problems actually stem from the off-manifold phenomenon associated with CFG, rather than the diffusion models themselves. More specifically, inspired by the recent advancements of diffusion model-based inverse problem solvers (DIS), we reformulate text-guidance as an inverse problem with a text-conditioned score matching loss and develop CFG++, a novel approach that tackles the off-manifold challenges inherent in traditional CFG. CFG++ features a surprisingly simple fix to CFG, yet it offers significant improvements, including better sample quality for text-to-image generation, invertibility, smaller guidance scales, reduced mode collapse, etc. Furthermore, CFG++ enables seamless interpolation between unconditional and conditional sampling at lower guidance scales, consistently outperforming traditional CFG at all scales. Moreover, CFG++ can be easily integrated into high-order diffusion solvers and naturally extends to distilled diffusion models. Experimental results confirm that our method significantly enhances performance in text-to-image generation, DDIM inversion, editing, and solving inverse problems, suggesting a wide-ranging impact and potential applications in various fields that utilize text guidance. Project Page: https://cfgpp-diffusion.github.io/.

Jaana Kettunen, Sally-Anne Barnes, Jenny Bimrose et al.

Abstract This article reports the findings from a phenomenographic study of career experts’ conceptions of systems development in lifelong guidance settings. The results show that conceptions of systems development in lifelong guidance varied from minimal, aspirational, strategic to systemic. By exploring the logical relationship between qualitatively different conceptions, it provides policymakers and other stakeholders with a way of holistically viewing the varying levels of lifelong guidance systems development. The matrix presented in this article may serve as a catalyst for reflection on crucial elements, such as legislation, leadership and cooperation, that have the potential to improve systems development in lifelong guidance.

Fredrik Hertzberg

Potentiality, actuality or exposure? The significance of recognition in career guidance counselling for newly arrived migrants. This article analyses the significance of recognition in career guidance counselling for newly arrived migrant youth. In line with the sociology of Herbert Blumer, the concept of recognition is employed as a sensitizing concept. In the article, several different aspects of recognition are highlighted, as recognition of prior learning, competencies, and experiences; educational and vocational aspirations; norms and values; the prospect of individual development, and the recognition of harsh and therefore noteworthy conditions for living. The study is based on an analysis of qualitative data from semi-structured interviews with fourteen career guidance counsellors that meet newly arrived migrants in the course of their daily work. The analysis draws attention to the fact that opportunities for development, as well as the recognition of particularly difficult and thus noteworthy life circumstances, are usually acknowledged without more extensive objections, but that the recognition of previously established knowledge, skills and experiences, study and career aspirations as well as norms and values can, according to the interviewees, lead to goal conflicts in study and career counselling work (although the relevance of a recognition approach is generally emphasised).

Gen Li, Xilong Jing, Tony Xing Tan et al.

Due to China???s three years of ???Zero Covid??? containment policy, the vast majority of vocational and technological education (VTE) high school students could not participate in work-based learning, had to take classes online, and often were isolated from their families. Although the impact of the unprecedented disruptions of the Covid-19 pandemic on students??? wellbeing has been well documented, little is known about the experiences of vocational and technical high school students whose vocational identity development depends heavily on work-based learning experiences. In the following study, we applied a risk and resilience framework to examine Chinese VTE student burnout following the exposure to the ???Zero Covid??? policy for their entire high school career. Specifically, we tested how variations in Chinese vocational and technical high school students??? perceived impact of the Covid-19 pandemic, career adaptability, psychological capital, academic performance, and parental career guidance were related to differences in their sense of burnout. Regression results showed that paternal education level and perceived impact of Covid-19 pandemic positively and significantly predicted burnout, while academic performance and parental career support negatively and significantly predicted burnout. Age, sex, internship participation, career adaptability and parental career interference and parental lack-of-engagement did not predict burnout. Overall, the variables accounted for 53.6% of the variance in student burnout.

Purwanto, Tjahyaningtijas Hapsari Peni Agustin, Paragas Jesse R. et al.

Clustering the distribution of student graduates is an approach used to analyze and understand the success of Vocational High School education programs in preparing graduates to enter the workforce or start their own businesses. The purpose of clustering is to evaluate the effectiveness of educational programs, identify entrepreneurial potential, formulate career planning, and develop entrepreneurial skills, all contributing to the fulfilment of Sustainable Development Goals (SDGs) related to quality education (SDG 4), decent work and economic growth (SDG 8), and industry, innovation, and infrastructure (SDG 9). Through this clustering, schools can evaluate the extent to which quality high school graduates achieve career or entrepreneurial success, supporting the objectives of SDG 4. This information helps in designing educational programs that are more in line with the needs of the job market, providing better career guidance to students, and promoting entrepreneurial skills among high school students, contributing to SDG 4 and SDG 8. Clustering the distribution of vocational high school students by working, continuing, and entrepreneurial status plays an important role in strengthening the link between education and the world of work, aligning with the aims of SDG 4 and SDG 8. Self-Organizing Map (SOM), as an Artificial Neural Network, assists in data clustering or mapping tasks, aiding in the discovery of patterns and trends within the vocational high school graduate population. The result of clustering using Z-Score and Min-Max normalization techniques is 5.31% and 3.98%, respectively, providing insights into the career and entrepreneurship trends and patterns of vocational high school students. This valuable information can be used for the development of educational programs, career guidance initiatives, and improved alignment between education and the needs of the world of work, ultimately contributing to the realization of SDGs 4, 8, and 9.

معصومه عامریان

ایجاد گلخانه برای تولید محصولات کشاورزی از راهبردهای مهم تضمین امنیت غذایی است. با این وجود توسعه‌ی گلخانه با مشکلات و موانع مختلفی مواجه است. لذا، هدف پژوهش حاضر بررسی مشکلات تولید و توسعه‌‌‌‌ی محصولات گلخانه‌ای شهرستان کرمانشاه می‌باشد. جامعه آماری شامل 100 گلخانه‌دار فعال در سال 1401 بود که تعداد 80 نفر از آن‌ها به‌عنوان نمونه‌ی آماری با استفاده از جدول کرجسی و مورگان انتخاب شدند. روش نمونه‌گیری از نوع تصادفی‌ساده بود. ابزار گردآوری داده‌ها پرسش‌نامه‌ی پژوهشگر ساخته بود که روایی و پایایی آن به‌ترتیب با استفاده از محاسبه‌ی نسبت روایی محتوا (CVR) و ضریب آلفای کرونباخ (81/0α≥) تأیید شد. داده‌های به‌دست آمده با روش‌های مانند آمار توصیفی (فراوانی، میانگین، انحراف‌معیار)، آمار استنباطی (تحلیل عاملی اکتشافی، u مان‌ویتنی و کروسکال والیس) با نرم‌افزار SPSS-20 بررسی شدند. یافته‌های تحلیل عاملی اکتشافی به‌روش واریماکس نشان داد که مشکلات تولید و توسعه‌ی محصولات گلخانه‌ای در شش عامل موانع اعتباری، موانع فنی و مشاوره‌ای، آموزشی و قانونی، تولید و صادرات، زیرساختی، و فروش قابل شناسایی هستند. این عامل‌ها در مجموع 98/64 درصد از واریانس متغیر وابسته را تبیین کردند. طبق نتایج به‌دست آمده از آزمون کروسکال والیس، اختلاف معنی‌داری (در سطح پنج درصد خطا) در دو گروه تحصیلات و نوع محصول تولیدی در گلخانه به‌ترتیب در دو عامل بازدارنده‌ی فنی و مشاوره‌ای و تولید و صادرات وجود دارد که افراد تحصیلکرده موانع فنی را مهم‌تر ارزیابی کردند. آزمون U من‌ویتنی در سطوح مختلف مالکیت معنی‌دار نبود.

Niall L. Williams, Nicholas Rewkowski, Jiasheng Li et al.

Haptic feedback is an important component of creating an immersive mixed reality experience. Traditionally, haptic forces are rendered in response to the user's interactions with the virtual environment. In this work, we explore the idea of rendering haptic forces in a proactive manner, with the explicit intention to influence the user's behavior through compelling haptic forces. To this end, we present a framework for active haptic guidance in mixed reality, using one or more robotic haptic proxies to influence user behavior and deliver a safer and more immersive virtual experience. We provide details on common challenges that need to be overcome when implementing active haptic guidance, and discuss example applications that show how active haptic guidance can be used to influence the user's behavior. Finally, we apply active haptic guidance to a virtual reality navigation problem, and conduct a user study that demonstrates how active haptic guidance creates a safer and more immersive experience for users.

Wayner Barrios, Mattia Soldan, Alberto Mario Ceballos-Arroyo et al.

The recent introduction of the large-scale, long-form MAD and Ego4D datasets has enabled researchers to investigate the performance of current state-of-the-art methods for video grounding in the long-form setup, with interesting findings: current grounding methods alone fail at tackling this challenging task and setup due to their inability to process long video sequences. In this paper, we propose a method for improving the performance of natural language grounding in long videos by identifying and pruning out non-describable windows. We design a guided grounding framework consisting of a Guidance Model and a base grounding model. The Guidance Model emphasizes describable windows, while the base grounding model analyzes short temporal windows to determine which segments accurately match a given language query. We offer two designs for the Guidance Model: Query-Agnostic and Query-Dependent, which balance efficiency and accuracy. Experiments demonstrate that our proposed method outperforms state-of-the-art models by 4.1% in MAD and 4.52% in Ego4D (NLQ), respectively. Code, data and MAD's audio features necessary to reproduce our experiments are available at: https://github.com/waybarrios/guidance-based-video-grounding.