Satya Krishna Gorti, Ilan Gofman, Zhaoyan Liu, Jiapeng Wu, Noël Vouitsis, Guangwei Yu, Jesse C. Cresswell, Rasa Hosseinzadeh. Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2025.

Tabular data is one of the most ubiquitous sources of information worldwide, spanning a wide variety of domains. This inherent heterogeneity has slowed the development of Tabular Foundation Models (TFMs) capable of fast generalization to unseen datasets. In-Context Learning (ICL) has recently emerged as a promising solution for TFMs, enabling dynamic adaptation to new tasks without additional tuning. While many studies have attempted to re-purpose large language models for tabular ICL, they have had limited success, so recent works have focused on developing tabular-specific foundation models. In this work, we propose an approach to combine ICL-based retrieval with self supervised learning to train tabular foundation models. We also investigate the utility of real vs. synthetic data for model pre-training, and show that real data can contain useful signal not easily captured in synthetic training. Specifically, we show that incorporating real data during the pre-training phase can lead to significantly faster training and better downstream generalization to unseen data. Our resulting model, **TabDPT**, achieves strong performance on both regression (CTR23) and classification (CC18) benchmarks. Importantly, we also demonstrate that with our pre-training procedure, scaling both model and data size leads to consistent performance improvements that follow power laws. This echoes scaling laws in LLMs and other foundation models, and suggests that large-scale TFMs can be achievable. We open-source our full pipeline: inference code including trained model weights can be found [here](https://github.com/layer6ai-labs/TabDPT-inference), and the training code to reproduce experiments can be found [here](https://github.com/layer6ai-labs/TabDPT-training).

Learning-based computer-generated holography (CGH) offers significant potential for holographic displays as a rapid hologram generation approach. However, as the target depth is fixed during network training, the generated holograms can only reconstruct images at a specific distance, posing a challenge for multi-depth displays. Here, we propose a depth-aware neural network for holographic displays. Our proposed method can generate high-quality holograms at any depth within a specified range by integrating a depth-aware module consisting of depth-embedding and self-attention blocks. Consequently, our network no longer needs to be retrained when the specific position of the object changes. Numerical reconstructions and optical experiments demonstrate that our method maintains high-quality performance within the training range and can understand the relationship between depth and holograms, reconstructing them clearly at untrained depths.

The goal of multimodal alignment is to learn a single latent space that is shared between multimodal inputs. The most powerful models in this space have been trained using massive datasets of paired inputs and large-scale computational resources, making them prohibitively expensive to train in many practical scenarios. We surmise that existing unimodal encoders pre-trained on large amounts of unimodal data should provide an effective bootstrap to create multimodal models from unimodal ones at much lower costs. We therefore propose FuseMix, a multimodal augmentation scheme that operates on the latent spaces of arbitrary pre-trained unimodal encoders. Using FuseMix for multimodal alignment, we achieve competitive performance - and in certain cases outperform state-of-the art methods - in both image-text and audio-text retrieval, with orders of magnitude less compute and data: for example, we outperform CLIP on the Flickr30K text-to-image retrieval task with ~600 × fewer GPU days and ~80 × fewer image-text pairs. Additionally, we show how our method can be applied to convert pre-trained text-to-image generative models into audio-to-image ones. Code is available at: https://github.com/layer6ai-labs/fusemix.

Tabular data is a pervasive modality spanning a wide range of domains, and this inherent diversity poses a considerable challenge for deep learning. Recent advancements using transformer-based in-context learning have shown promise on smaller and less complex tabular datasets, but have struggled to scale to larger and more complex ones. To address this limitation, we propose a combination of retrieval and fine-tuning: we can adapt the transformer to a local subset of the data by collecting nearest neighbours, and then perform task-specific fine-tuning with this retrieved set of neighbours in context. Using TabPFN as the base model -- currently the best tabular in-context learner -- and applying our retrieval and fine-tuning scheme on top results in what we call a locally-calibrated PFN, or LoCalPFN. We conduct extensive evaluation on 95 datasets curated by TabZilla from OpenML, upon which we establish a new state-of-the-art with LoCalPFN -- even with respect to tuned tree-based models. Notably, we show a significant boost in performance compared to the base in-context model, demonstrating the efficacy of our approach and advancing the frontier of deep learning in tabular data.

In text-video retrieval, the objective is to learn a cross-modal similarity function between a text and a video that ranks relevant text-video pairs higher than irrelevant pairs. However, videos inherently express a much wider gamut of information than texts. Instead, texts often capture sub-regions of entire videos and are most semantically similar to certain frames within videos. Therefore, for a given text, a retrieval model should focus on the text's most semantically similar video sub-regions to make a more relevant comparison. Yet, most existing works aggregate entire videos with-out directly considering text. Common text-agnostic ag-gregations schemes include mean-pooling or self-attention over the frames, but these are likely to encode misleading vi-sual information not described in the given text. To address this, we propose a cross-modal attention model called X-Pool that reasons between a text and the frames of a video. Our core mechanism is a scaled dot product attention for a text to attend to its most semantically similar frames. We then generate an aggregated video representation conditioned on the text's attention weights over the frames. We evaluate our method on three benchmark datasets of MSR-VTT, MSVD and LSMDC, achieving new state-of-the-art re-sults by up to 12% in relative improvement in Recall@ 1. Our findings thereby highlight the importance of joint text-video reasoning to extract important visual cues according to text. Full code and demo can be found at: layer6ai-labs.github.iolxpooll.

Localizing actions in video is a core task in computer vision. The weakly supervised temporal localization problem investigates whether this task can be adequately solved with only video-level labels, significantly reducing the amount of expensive and error-prone annotation that is required. A common approach is to train a frame-level classifier where frames with the highest class probability are selected to make a video-level prediction. Frame-level activations are then used for localization. However, the absence of frame-level annotations cause the classifier to impart class bias on every frame. To address this, we propose the Action Selection Learning (ASL) approach to capture the general concept of action, a property we refer to as "actionness". Under ASL, the model is trained with a novel class-agnostic task to predict which frames will be selected by the classifier. Empirically, we show that ASL outperforms leading baselines on two popular bench-marks THUMOS-14 and ActivityNet-1.2, with 10.3% and 5.7% relative improvement respectively. We further analyze the properties of ASL and demonstrate the importance of actionness. Full code for this work is available here: https://github.com/layer6ai-labs/ASL.

Scene graph generation is an important task in computer vision aimed at improving the semantic understanding of the visual world. In this task, the model needs to detect objects and predict visual relationships between them. Most of the existing models predict relationships in parallel assuming their independence. While there are different ways to capture these dependencies, we explore a conditional approach motivated by the sequence-to-sequence (Seq2Seq) formalism. Different from the previous research, our proposed model predicts visual relationships one at a time in an autoregressive manner by explicitly conditioning on the already predicted relationships. Drawing from translation models in NLP, we propose an encoder-decoder model built using Transformers where the encoder captures global context and long range interactions. The decoder then makes sequential predictions by conditioning on the scene graph constructed so far. In addition, we introduce a novel reinforcement learning-based training strategy tailored to Seq2Seq scene graph generation. By using a self-critical policy gradient training approach with Monte Carlo search we directly optimize for the (mean) recall metrics and bridge the gap between training and evaluation. Experimental results on two public benchmark datasets demonstrate that our Seq2Seq learning approach achieves strong empirical performance, outperforming previous state-of-the-art, while remaining efficient in terms of training and inference time. Full code for this work is available here: https://github.com/layer6ai-labs/SGG-Seq2Seq.

We propose a novel graph-based approach for image retrieval. Given a nearest neighbor graph produced by the global descriptor model, we traverse it by alternating between exploit and explore steps. The exploit step maximally utilizes the immediate neighborhood of each vertex, while the explore step traverses vertices that are farther away in the descriptor space. By combining these two steps we can better capture the underlying image manifold, and successfully retrieve relevant images that are visually dissimilar to the query. Our traversal algorithm is conceptually simple, has few tunable parameters and can be implemented with basic data structures. This enables fast real-time inference for previously unseen queries with minimal memory overhead. Despite relative simplicity, we show highly competitive results on multiple public benchmarks, including the largest image retrieval dataset that is currently publicly available. Full code for this work is available here: https://github.com/layer6ai-labs/egt.

Despite recent progress in computer vision, image retrieval remains a challenging open problem. Numerous variations such as view angle, lighting and occlusion make it difficult to design models that are both robust and efficient. Many leading methods traverse the nearest neighbor graph to exploit higher order neighbor information and uncover the highly complex underlying manifold. In this work we propose a different approach where we leverage graph convolutional networks to directly encode neighbor information into image descriptors. We further leverage ideas from clustering and manifold learning, and introduce an unsupervised loss based on pairwise separation of image similarities. Empirically, we demonstrate that our model is able to successfully learn a new descriptor space that significantly improves retrieval accuracy, while still allowing efficient inner product inference. Experiments on five public benchmarks show highly competitive performance with up to 24\% relative improvement in mAP over leading baselines. Full code for this work is available here: https://github.com/layer6ai-labs/GSS.

Latent models have become the default choice for recommender systems due to their performance and scalability. However, research in this area has primarily focused on modeling user-item interactions, and few latent models have been developed for cold start. Deep learning has recently achieved remarkable success showing excellent results for diverse input types. Inspired by these results we propose a neural network based latent model called DropoutNet to address the cold start problem in recommender systems. Unlike existing approaches that incorporate additional content-based objective terms, we instead focus on the optimization and show that neural network models can be explicitly trained for cold start through dropout. Our model can be applied on top of any existing latent model effectively providing cold start capabilities, and full power of deep architectures. Empirically we demonstrate state-of-the-art accuracy on publicly available benchmarks. Code is available at https://github.com/layer6ai-labs/DropoutNet.

In many collaborative filtering (CF) applications, latent approaches are the preferred model choice due to their ability to generate real-time recommendations efficiently. However, the majority of existing latent models are not designed for implicit binary feedback (views, clicks, plays etc.) and perform poorly on data of this type. Developing accurate models from implicit feedback is becoming increasingly important in CF since implicit feedback can often be collected at lower cost and in much larger quantities than explicit preferences. The need for accurate latent models for implicit data was further emphasized by the recently conducted Million Song Dataset Challenge organized by Kaggle. In this challenge, the results for the best latent model were orders of magnitude worse than neighbor-based approaches, and all the top performing teams exclusively used neighbor-based models. We address this problem and propose a new latent approach for binary feedback in CF. In our model, neighborhood similarity information is used to guide latent factorization and derive accurate latent representations. We show that even with simple factorization methods like SVD, our approach outperforms existing models and produces state-of-the-art results.