arXiv subjects

Oswald Lanz

Publications and source records attributed to Oswald Lanz.

At least 19 recordsLinked to original sources

arXiv2026

Investigating Knowledge Transfer Across Interactive Dialogue Games

Dialogue games represent a challenging setting where complex cognitive skills are required to accomplish tasks while coordinating with other players. Considering that language represents an interface for both understanding the game rules and executing actions, it is reasonable to assume that training on a specific language game will enhance specific capabilities that might be relevant for other tasks as well. Motivated by this rationale, in this paper, we investigate how knowledge transfers across different dialogue games. We study transferability by finetuning LLM models on games from the clembench suite (Chalamalasetti et al., 2023) and performing two analyses: i) we derive a task-transferability graph using a binary integer optimization program from Zamir et al. (2018), using task performance as the main metric; and ii) we compute task vectors (Ilharco et al., 2022) for each game to study similarities across finetuned models and their task transferability. In our first analysis, we find that some games benefit more from transfer than finetuning, and that the visuospatial family (e.g., exploration games) transfers best. With our task vector analysis instead, we find that similarity-based approaches capture game-role relationships but almost no transferability patterns, suggesting that more complex metrics are required.

cs.CL↗

arXiv2026

Divide, Deliberate, Decide: A Multi-Agent Framework for Fine-Grained Egocentric Action Recognition

Fine-grained action recognition in egocentric video is challenging for Vision-Language Models (VLMs): actions often differ only in small visual cues, and a single model tends to be biased toward a subset of these cues. We propose Divide, Deliberate, Decide, a fully-local, zero-shot multi-agent framework in which (i) a VLM orchestrator chunks the video and proposes a top-k candidate label list per segment, (ii) an ensemble of heterogeneous VLM specialists, drawn from different open model families, engages in a structured deliberation that includes a peer-consultation round of questions, and (iii) agent rankings are aggregated with a Borda count and the orchestrator re-ranks its own prediction in light of the specialists' evidence. The entire pipeline runs locally with no fine-tuning. Experiments show that our method positively improves zero-shot action recognition performance over the baseline, highlighting the influence of a heterogeneous deliberation step, showing that the gain stems from decorrelated model priors rather than from additional compute.

cs.CV↗

arXiv2026

Action-Guided Attention for Video Action Anticipation

Anticipating future actions in videos is challenging, as the observed frames provide only evidence of past activities, requiring the inference of latent intentions to predict upcoming actions. Existing transformer-based approaches, which rely on dot-product attention over pixel representations, often lack the high-level semantics necessary to model video sequences for effective action anticipation. As a result, these methods tend to overfit to explicit visual cues present in the past frames, limiting their ability to capture underlying intentions and degrading generalization to unseen samples. To address this, we propose Action-Guided Attention (AGA), an attention mechanism that explicitly leverages predicted action sequences as queries and keys to guide sequence modeling. Our approach fosters the attention module to emphasize relevant moments from the past based on the upcoming activity and combine this information with the current frame embedding via a dedicated gating function. The design of AGA enables post-training analysis of the knowledge discovered from the training set. Experiments on the widely adopted EPIC-Kitchens-100 benchmark demonstrate that AGA generalizes well from validation to unseen test sets. Post-training analysis can further examine the action dependencies captured by the model and the counterfactual evidence it has internalized, offering transparent and interpretable insights into its anticipative predictions.

cs.CV↗

arXiv2025

Gate-Shift-Pose: Enhancing Action Recognition in Sports with Skeleton Information

This paper introduces Gate-Shift-Pose, an enhanced version of Gate-Shift-Fuse networks, designed for athlete fall classification in figure skating by integrating skeleton pose data alongside RGB frames. We evaluate two fusion strategies: early-fusion, which combines RGB frames with Gaussian heatmaps of pose keypoints at the input stage, and late-fusion, which employs a multi-stream architecture with attention mechanisms to combine RGB and pose features. Experiments on the FR-FS dataset demonstrate that Gate-Shift-Pose significantly outperforms the RGB-only baseline, improving accuracy by up to 40% with ResNet18 and 20% with ResNet50. Early-fusion achieves the highest accuracy (98.08%) with ResNet50, leveraging the model's capacity for effective multimodal integration, while late-fusion is better suited for lighter backbones like ResNet18. These results highlight the potential of multimodal architectures for sports action recognition and the critical role of skeleton pose information in capturing complex motion patterns. Visit the project page at https://edowhite.github.io/Gate-Shift-Pose

cs.CV↗

arXiv2025

Movie Facts and Fibs (MF$^2$): A Benchmark for Long Movie Understanding

Despite recent progress in vision-language models (VLMs), holistic understanding of long-form video content remains a significant challenge, partly due to limitations in current benchmarks. Many focus on peripheral, ``needle-in-a-haystack'' details, encouraging context-insensitive retrieval over deep comprehension. Others rely on large-scale, semi-automatically generated questions (often produced by language models themselves) that are easier for models to answer but fail to reflect genuine understanding. In this paper, we introduce MF$^2$, a new benchmark for evaluating whether models can comprehend, consolidate, and recall key narrative information from full-length movies (50-170 minutes long). MF$^2$ includes over 50 full-length, open-licensed movies, each paired with manually constructed sets of claim pairs -- one true (fact) and one plausible but false (fib), totalling over 850 pairs. These claims target core narrative elements such as character motivations and emotions, causal chains, and event order, and refer to memorable moments that humans can recall without rewatching the movie. Instead of multiple-choice formats, we adopt a binary claim evaluation protocol: for each pair, models must correctly identify both the true and false claims. This reduces biases like answer ordering and enables a more precise assessment of reasoning. Our experiments demonstrate that both open-weight and closed state-of-the-art models fall well short of human performance, underscoring the relative ease of the task for humans and their superior ability to retain and reason over critical narrative information -- an ability current VLMs lack.

cs.CV↗

arXiv2025

L-SWAG: Layer-Sample Wise Activation with Gradients information for Zero-Shot NAS on Vision Transformers

Training-free Neural Architecture Search (NAS) efficiently identifies high-performing neural networks using zero-cost (ZC) proxies. Unlike multi-shot and one-shot NAS approaches, ZC-NAS is both (i) time-efficient, eliminating the need for model training, and (ii) interpretable, with proxy designs often theoretically grounded. Despite rapid developments in the field, current SOTA ZC proxies are typically constrained to well-established convolutional search spaces. With the rise of Large Language Models shaping the future of deep learning, this work extends ZC proxy applicability to Vision Transformers (ViTs). We present a new benchmark using the Autoformer search space evaluated on 6 distinct tasks and propose Layer-Sample Wise Activation with Gradients information (L-SWAG), a novel, generalizable metric that characterizes both convolutional and transformer architectures across 14 tasks. Additionally, previous works highlighted how different proxies contain complementary information, motivating the need for a ML model to identify useful combinations. To further enhance ZC-NAS, we therefore introduce LIBRA-NAS (Low Information gain and Bias Re-Alignment), a method that strategically combines proxies to best represent a specific benchmark. Integrated into the NAS search, LIBRA-NAS outperforms evolution and gradient-based NAS techniques by identifying an architecture with a 17.0% test error on ImageNet1k in just 0.1 GPU days.

cs.CV↗

arXiv2024

GRASP-GCN: Graph-Shape Prioritization for Neural Architecture Search under Distribution Shifts

Neural Architecture Search (NAS) methods have shown to output networks that largely outperform human-designed networks. However, conventional NAS methods have mostly tackled the single dataset scenario, incuring in a large computational cost as the procedure has to be run from scratch for every new dataset. In this work, we focus on predictor-based algorithms and propose a simple and efficient way of improving their prediction performance when dealing with data distribution shifts. We exploit the Kronecker-product on the randomly wired search-space and create a small NAS benchmark composed of networks trained over four different datasets. To improve the generalization abilities, we propose GRASP-GCN, a ranking Graph Convolutional Network that takes as additional input the shape of the layers of the neural networks. GRASP-GCN is trained with the not-at-convergence accuracies, and improves the state-of-the-art of 3.3 % for Cifar-10 and increasing moreover the generalization abilities under data distribution shift.

cs.CV↗

arXiv2024

Your Image is My Video: Reshaping the Receptive Field via Image-To-Video Differentiable AutoAugmentation and Fusion

The landscape of deep learning research is moving towards innovative strategies to harness the true potential of data. Traditionally, emphasis has been on scaling model architectures, resulting in large and complex neural networks, which can be difficult to train with limited computational resources. However, independently of the model size, data quality (i.e. amount and variability) is still a major factor that affects model generalization. In this work, we propose a novel technique to exploit available data through the use of automatic data augmentation for the tasks of image classification and semantic segmentation. We introduce the first Differentiable Augmentation Search method (DAS) to generate variations of images that can be processed as videos. Compared to previous approaches, DAS is extremely fast and flexible, allowing the search on very large search spaces in less than a GPU day. Our intuition is that the increased receptive field in the temporal dimension provided by DAS could lead to benefits also to the spatial receptive field. More specifically, we leverage DAS to guide the reshaping of the spatial receptive field by selecting task-dependant transformations. As a result, compared to standard augmentation alternatives, we improve in terms of accuracy on ImageNet, Cifar10, Cifar100, Tiny-ImageNet, Pascal-VOC-2012 and CityScapes datasets when plugging-in our DAS over different light-weight video backbones.

cs.CV↗

arXiv2023

Gate-Shift-Fuse for Video Action Recognition

Convolutional Neural Networks are the de facto models for image recognition. However 3D CNNs, the straight forward extension of 2D CNNs for video recognition, have not achieved the same success on standard action recognition benchmarks. One of the main reasons for this reduced performance of 3D CNNs is the increased computational complexity requiring large scale annotated datasets to train them in scale. 3D kernel factorization approaches have been proposed to reduce the complexity of 3D CNNs. Existing kernel factorization approaches follow hand-designed and hard-wired techniques. In this paper we propose Gate-Shift-Fuse (GSF), a novel spatio-temporal feature extraction module which controls interactions in spatio-temporal decomposition and learns to adaptively route features through time and combine them in a data dependent manner. GSF leverages grouped spatial gating to decompose input tensor and channel weighting to fuse the decomposed tensors. GSF can be inserted into existing 2D CNNs to convert them into an efficient and high performing spatio-temporal feature extractor, with negligible parameter and compute overhead. We perform an extensive analysis of GSF using two popular 2D CNN families and achieve state-of-the-art or competitive performance on five standard action recognition benchmarks.

cs.CV↗

arXiv2023

Inductive Attention for Video Action Anticipation

Anticipating future actions based on spatiotemporal observations is essential in video understanding and predictive computer vision. Moreover, a model capable of anticipating the future has important applications, it can benefit precautionary systems to react before an event occurs. However, unlike in the action recognition task, future information is inaccessible at observation time -- a model cannot directly map the video frames to the target action to solve the anticipation task. Instead, the temporal inference is required to associate the relevant evidence with possible future actions. Consequently, existing solutions based on the action recognition models are only suboptimal. Recently, researchers proposed extending the observation window to capture longer pre-action profiles from past moments and leveraging attention to retrieve the subtle evidence to improve the anticipation predictions. However, existing attention designs typically use frame inputs as the query which is suboptimal, as a video frame only weakly connects to the future action. To this end, we propose an inductive attention model, dubbed IAM, which leverages the current prediction priors as the query to infer future action and can efficiently process the long video content. Furthermore, our method considers the uncertainty of the future via the many-to-many association in the attention design. As a result, IAM consistently outperforms the state-of-the-art anticipation models on multiple large-scale egocentric video datasets while using significantly fewer model parameters.

cs.CV↗

arXiv2022

A Feature-space Multimodal Data Augmentation Technique for Text-video Retrieval

Every hour, huge amounts of visual contents are posted on social media and user-generated content platforms. To find relevant videos by means of a natural language query, text-video retrieval methods have received increased attention over the past few years. Data augmentation techniques were introduced to increase the performance on unseen test examples by creating new training samples with the application of semantics-preserving techniques, such as color space or geometric transformations on images. Yet, these techniques are usually applied on raw data, leading to more resource-demanding solutions and also requiring the shareability of the raw data, which may not always be true, e.g. copyright issues with clips from movies or TV series. To address this shortcoming, we propose a multimodal data augmentation technique which works in the feature space and creates new videos and captions by mixing semantically similar samples. We experiment our solution on a large scale public dataset, EPIC-Kitchens-100, and achieve considerable improvements over a baseline method, improved state-of-the-art performance, while at the same time performing multiple ablation studies. We release code and pretrained models on Github at https://github.com/aranciokov/FSMMDA_VideoRetrieval.

cs.CV↗

arXiv2022

NVIDIA-UNIBZ Submission for EPIC-KITCHENS-100 Action Anticipation Challenge 2022

In this report, we describe the technical details of our submission for the EPIC-Kitchen-100 action anticipation challenge. Our modelings, the higher-order recurrent space-time transformer and the message-passing neural network with edge learning, are both recurrent-based architectures which observe only 2.5 seconds inference context to form the action anticipation prediction. By averaging the prediction scores from a set of models compiled with our proposed training pipeline, we achieved strong performance on the test set, which is 19.61% overall mean top-5 recall, recorded as second place on the public leaderboard.

cs.CV↗

arXiv2022

UniUD-FBK-UB-UniBZ Submission to the EPIC-Kitchens-100 Multi-Instance Retrieval Challenge 2022

This report presents the technical details of our submission to the EPIC-Kitchens-100 Multi-Instance Retrieval Challenge 2022. To participate in the challenge, we designed an ensemble consisting of different models trained with two recently developed relevance-augmented versions of the widely used triplet loss. Our submission, visible on the public leaderboard, obtains an average score of 61.02% nDCG and 49.77% mAP.

cs.CV↗

arXiv2022

Unified Recurrence Modeling for Video Action Anticipation

Forecasting future events based on evidence of current conditions is an innate skill of human beings, and key for predicting the outcome of any decision making. In artificial vision for example, we would like to predict the next human action before it happens, without observing the future video frames associated to it. Computer vision models for action anticipation are expected to collect the subtle evidence in the preamble of the target actions. In prior studies recurrence modeling often leads to better performance, the strong temporal inference is assumed to be a key element for reasonable prediction. To this end, we propose a unified recurrence modeling for video action anticipation via message passing framework. The information flow in space-time can be described by the interaction between vertices and edges, and the changes of vertices for each incoming frame reflects the underlying dynamics. Our model leverages self-attention as the building blocks for each of the message passing functions. In addition, we introduce different edge learning strategies that can be end-to-end optimized to gain better flexibility for the connectivity between vertices. Our experimental results demonstrate that our proposed method outperforms previous works on the large-scale EPIC-Kitchen dataset.

cs.CV↗

arXiv2022

Relevance-based Margin for Contrastively-trained Video Retrieval Models

Video retrieval using natural language queries has attracted increasing interest due to its relevance in real-world applications, from intelligent access in private media galleries to web-scale video search. Learning the cross-similarity of video and text in a joint embedding space is the dominant approach. To do so, a contrastive loss is usually employed because it organizes the embedding space by putting similar items close and dissimilar items far. This framework leads to competitive recall rates, as they solely focus on the rank of the groundtruth items. Yet, assessing the quality of the ranking list is of utmost importance when considering intelligent retrieval systems, since multiple items may share similar semantics, hence a high relevance. Moreover, the aforementioned framework uses a fixed margin to separate similar and dissimilar items, treating all non-groundtruth items as equally irrelevant. In this paper we propose to use a variable margin: we argue that varying the margin used during training based on how much relevant an item is to a given query, i.e. a relevance-based margin, easily improves the quality of the ranking lists measured through nDCG and mAP. We demonstrate the advantages of our technique using different models on EPIC-Kitchens-100 and YouCook2. We show that even if we carefully tuned the fixed margin, our technique (which does not have the margin as a hyper-parameter) would still achieve better performance. Finally, extensive ablation studies and qualitative analysis support the robustness of our approach. Code will be released at \url{https://github.com/aranciokov/RelevanceMargin-ICMR22}.

cs.CV↗

arXiv2022

Learning video retrieval models with relevance-aware online mining

Due to the amount of videos and related captions uploaded every hour, deep learning-based solutions for cross-modal video retrieval are attracting more and more attention. A typical approach consists in learning a joint text-video embedding space, where the similarity of a video and its associated caption is maximized, whereas a lower similarity is enforced with all the other captions, called negatives. This approach assumes that only the video and caption pairs in the dataset are valid, but different captions - positives - may also describe its visual contents, hence some of them may be wrongly penalized. To address this shortcoming, we propose the Relevance-Aware Negatives and Positives mining (RANP) which, based on the semantics of the negatives, improves their selection while also increasing the similarity of other valid positives. We explore the influence of these techniques on two video-text datasets: EPIC-Kitchens-100 and MSR-VTT. By using the proposed techniques, we achieve considerable improvements in terms of nDCG and mAP, leading to state-of-the-art results, e.g. +5.3% nDCG and +3.0% mAP on EPIC-Kitchens-100. We share code and pretrained models at \url{https://github.com/aranciokov/ranp}.

cs.CV↗

arXiv2021

SAIC_Cambridge-HuPBA-FBK Submission to the EPIC-Kitchens-100 Action Recognition Challenge 2021

This report presents the technical details of our submission to the EPIC-Kitchens-100 Action Recognition Challenge 2021. To participate in the challenge we deployed spatio-temporal feature extraction and aggregation models we have developed recently: GSF and XViT. GSF is an efficient spatio-temporal feature extracting module that can be plugged into 2D CNNs for video action recognition. XViT is a convolution free video feature extractor based on transformer architecture. We design an ensemble of GSF and XViT model families with different backbones and pretraining to generate the prediction scores. Our submission, visible on the public leaderboard, achieved a top-1 action recognition accuracy of 44.82%, using only RGB.

cs.CV↗

arXiv2021

Higher Order Recurrent Space-Time Transformer for Video Action Prediction

Endowing visual agents with predictive capability is a key step towards video intelligence at scale. The predominant modeling paradigm for this is sequence learning, mostly implemented through LSTMs. Feed-forward Transformer architectures have replaced recurrent model designs in ML applications of language processing and also partly in computer vision. In this paper we investigate on the competitiveness of Transformer-style architectures for video predictive tasks. To do so we propose HORST, a novel higher order recurrent layer design whose core element is a spatial-temporal decomposition of self-attention for video. HORST achieves state of the art competitive performance on Something-Something early action recognition and EPIC-Kitchens action anticipation, showing evidence of predictive capability that we attribute to our recurrent higher order design of self-attention.

cs.CV↗