Oncel Tuzel

ICLR Conference 2025 Conference Paper

GSM-Symbolic: Understanding the Limitations of Mathematical Reasoning in Large Language Models

• Iman Mirzadeh
• Keivan Alizadeh
• Hooman Shahrokhi
• Oncel Tuzel
• Samy Bengio
• Mehrdad Farajtabar

Recent advancements in Large Language Models (LLMs) have sparked interest in their mathematical reasoning capabilities. While performance on the widely popular GSM8K benchmark has improved, questions remain about whether reported evaluation metrics are reliable, and reasoning abilities of LLMs have advanced. To overcome the limitations of existing evaluations, we introduce GSM-Symbolic, an improved benchmark created from symbolic templates that allow for the generation of a diverse set of questions. GSM-Symbolic enables more controllable evaluations, providing key insights and more reliable metrics for measuring the reasoning capabilities of models. Our findings reveal that LLMs exhibit noticeable variance when responding to different instantiations of the same question. Specifically, the performance of models declines when only the numerical values in the question are altered in the GSM-Symbolic benchmark. Furthermore, we investigate the fragility of mathematical reasoning in these models and demonstrate that their performance significantly deteriorates as the number of clauses in a question increases. We hypothesize that this decline is due to the fact that current LLMs are not capable of genuine logical reasoning; instead, they attempt to replicate the reasoning steps observed in their training data. When we add a single clause that appears relevant to the question, we observe significant performance drops (up to 65%) across all state-of-the-art models, even though the added clause does not contribute to the reasoning chain needed to reach the final answer. Overall, our work provides a more nuanced understanding of LLMs' capabilities and limitations in mathematical reasoning.

TMLR Journal 2025 Journal Article

MobileCLIP2: Improving Multi-Modal Reinforced Training

• Fartash Faghri
• Pavan Kumar Anasosalu Vasu
• Cem Koc
• Vaishaal Shankar
• Alexander T Toshev
• Oncel Tuzel
• Hadi Pouransari

Foundation image-text models such as CLIP with zero-shot capabilities enable a wide array of applications. MobileCLIP is a recent family of image-text models at 3-15ms latency and 50-150M parameters with state-of-the-art zero-shot accuracy. The main ingredients in MobileCLIP were its low-latency and light architectures and a novel multi-modal reinforced training that made knowledge distillation from multiple caption-generators and CLIP teachers efficient, scalable, and reproducible. In this paper, we improve the multi-modal reinforced training of MobileCLIP through: 1) better CLIP teacher ensembles trained on the DFN dataset, 2) improved captioner teachers trained on the DFN dataset and fine-tuned on a diverse selection of high-quality image-caption datasets. We discover new insights through ablations such as the importance of temperature tuning in contrastive knowledge distillation, the effectiveness of caption-generator fine-tuning for caption diversity, and the additive improvement from combining synthetic captions generated by multiple models. We train a new family of models called MobileCLIP2 and achieve state-of-the-art ImageNet-1k zero-shot accuracies at low latencies. In particular, we observe 2.2% improvement in ImageNet-1k accuracy for MobileCLIP2-B compared with MobileCLIP-B architecture. Notably, MobileCLIP2-S4 matches the zero-shot accuracy of SigLIP-SO400M/14 on ImageNet-1k while being 2× smaller and improves on DFN ViT-L/14 at 2.5× lower latency. We release our pretrained models and the data generation code. The data generation code makes it easy to create new reinforced datasets with arbitrary teachers using distributed scalable processing.

ICML Conference 2025 Conference Paper

Proxy-FDA: Proxy-based Feature Distribution Alignment for Fine-tuning Vision Foundation Models without Forgetting

• Chen Huang 0001
• Skyler Seto
• Hadi Pouransari
• Mehrdad Farajtabar
• Raviteja Vemulapalli
• Fartash Faghri
• Oncel Tuzel
• Barry-John Theobald

Vision foundation models pre-trained on massive data encode rich representations of real-world concepts, which can be adapted to downstream tasks by fine-tuning. However, fine-tuning foundation models on one task often leads to the issue of concept forgetting on other tasks. Recent methods of robust fine-tuning aim to mitigate forgetting of prior knowledge without affecting the fine-tuning performance. Knowledge is often preserved by matching the original and fine-tuned model weights or feature pairs. However, such point-wise matching can be too strong, without explicit awareness of the feature neighborhood structures that encode rich knowledge as well. We propose a novel regularization method Proxy-FDA that explicitly preserves the structural knowledge in feature space. Proxy-FDA performs Feature Distribution Alignment (using nearest neighbor graphs) between the pre-trained and fine-tuned feature spaces, and the alignment is further improved by informative proxies that are generated dynamically to increase data diversity. Experiments show that Proxy-FDA significantly reduces concept forgetting during fine-tuning, and we find a strong correlation between forgetting and a distributional distance metric (in comparison to L2 distance). We further demonstrate Proxy-FDA’s benefits in various fine-tuning settings (end-to-end, few-shot and continual tuning) and across different tasks like image classification, captioning and VQA.

TMLR Journal 2024 Journal Article

CLIP meets Model Zoo Experts: Pseudo-Supervision for Visual Enhancement

• Mohammadreza Salehi
• Mehrdad Farajtabar
• Maxwell Horton
• Fartash Faghri
• Hadi Pouransari
• Raviteja Vemulapalli
• Oncel Tuzel
• Ali Farhadi

Contrastive language image pretraining (CLIP) is a standard method for training vision-language models. While CLIP is scalable, promptable, and robust to distribution shifts on image classification tasks, it lacks object localization capabilities. This paper studies the following question: Can we augment CLIP training with task-specific vision models from model zoos to improve its visual representations? Towards this end, we leverage open-source task-specific vision models to generate pseudo-labels for an uncurated and noisy image-text dataset. Subsequently, we train CLIP models on these pseudo-labels in addition to the contrastive training on image and text pairs. This simple setup shows substantial improvements of up to 16.3% across different vision tasks, including segmentation, detection, depth estimation, and surface normal estimation. Importantly, these enhancements are achieved without compromising CLIP's existing capabilities, including its proficiency in promptable zero-shot classification.

NeurIPS Conference 2024 Conference Paper

Dataset Decomposition: Faster LLM Training with Variable Sequence Length Curriculum

• Hadi Pouransari
• Chun-Liang Li
• Jen-Hao R. Chang
• Pavan K. Vasu
• Cem Koc
• Vaishaal Shankar
• Oncel Tuzel

Large language models (LLMs) are commonly trained on datasets consisting of fixed-length token sequences. These datasets are created by randomly concatenating documents of various lengths and then chunking them into sequences of a predetermined target length (concat-and-chunk). Recent attention implementations mask cross-document attention, reducing the effective length of a chunk of tokens. Additionally, training on long sequences becomes computationally prohibitive due to the quadratic cost of attention. In this study, we introduce dataset decomposition, a novel variable sequence length training technique, to tackle these challenges. We decompose a dataset into a union of buckets, each containing sequences of the same size extracted from a unique document. During training, we use variable sequence length and batch-size, sampling simultaneously from all buckets with a curriculum. In contrast to the concat-and-chunk baseline, which incurs a fixed attention cost at every step of training, our proposed method incurs a computational cost proportional to the actual document lengths at each step, resulting in significant savings in training time. We train an 8k context-length 1B model at the same cost as a 2k context-length model trained with the baseline approach. Experiments on a web-scale corpus demonstrate that our approach significantly enhances performance on standard language evaluations and long-context benchmarks, reaching target accuracy with up to 6x faster training compared to the baseline. Our method not only enables efficient pretraining on long sequences but also scales effectively with dataset size. Lastly, we shed light on a critical yet less studied aspect of training large language models: the distribution and curriculum of sequence lengths, which results in a non-negligible difference in performance.

ICML Conference 2024 Conference Paper

Knowledge Transfer from Vision Foundation Models for Efficient Training of Small Task-specific Models

• Raviteja Vemulapalli
• Hadi Pouransari
• Fartash Faghri
• Sachin Mehta
• Mehrdad Farajtabar
• Mohammad Rastegari
• Oncel Tuzel

Vision Foundation Models (VFMs) pretrained on massive datasets exhibit impressive performance on various downstream tasks, especially with limited labeled target data. However, due to their high inference compute cost, these models cannot be deployed for many real-world applications. Motivated by this, we ask the following important question, "How can we leverage the knowledge from a large VFM to train a small task-specific model for a new target task with limited labeled training data? ", and propose a simple task-oriented knowledge transfer approach as a highly effective solution to this problem. Our experimental results on five target tasks show that the proposed approach outperforms task-agnostic VFM distillation, web-scale CLIP pretraining, supervised ImageNet pretraining, and self-supervised DINO pretraining by up to 11. 6%, 22. 1%, 13. 7%, and 29. 8%, respectively. Furthermore, the proposed approach also demonstrates up to 9x, 4x and 15x reduction in pretraining compute cost when compared to task-agnostic VFM distillation, ImageNet pretraining and DINO pretraining, respectively, while outperforming them. We also show that the dataset used for transferring knowledge has a significant effect on the final target task performance, and introduce a retrieval-augmented knowledge transfer strategy that uses web-scale image retrieval to curate effective transfer sets.

ICLR Conference 2024 Conference Paper

ReLU Strikes Back: Exploiting Activation Sparsity in Large Language Models

• Iman Mirzadeh
• Keivan Alizadeh
• Sachin Mehta
• Carlo C. del Mundo
• Oncel Tuzel
• Golnoosh Samei
• Mohammad Rastegari
• Mehrdad Farajtabar

Large Language Models (LLMs) with billions of parameters have drastically transformed AI applications. However, their demanding computation during inference has raised significant challenges for deployment on resource-constrained devices. Despite recent trends favoring alternative activation functions such as GELU or SiLU, known for increased computation, this study strongly advocates for reinstating ReLU activation in LLMs. We demonstrate that using the ReLU activation function has a negligible impact on convergence and performance while significantly reducing computation and weight transfer. This reduction is particularly valuable during the memory-bound inference step, where efficiency is paramount. Exploring sparsity patterns in ReLU-based LLMs, we unveil the reutilization of activated neurons for generating new tokens and leveraging these insights, we propose practical strategies to substantially reduce LLM inference computation up to three times, using ReLU activations with minimal performance trade-offs.

ICLR Conference 2024 Conference Paper

TiC-CLIP: Continual Training of CLIP Models

• Saurabh Garg
• Mehrdad Farajtabar
• Hadi Pouransari
• Raviteja Vemulapalli
• Sachin Mehta
• Oncel Tuzel
• Vaishaal Shankar
• Fartash Faghri

Keeping large foundation models up to date on latest data is inherently expensive. To avoid the prohibitive costs of constantly retraining, it is imperative to continually train these models. This problem is exacerbated by the lack of any large scale continual learning benchmarks or baselines. We introduce the first set of web-scale Time-Continual (TiC) benchmarks for training vision-language models: TiC-DataComp, TiC-YFCC, and TiC-Redcaps. TiC-DataComp, our largest dataset, contains over 12.7B timestamped image-text pairs spanning 9 years (2014-2022). We first use our benchmarks to curate various dynamic evaluations to measure temporal robustness of existing models. We show OpenAI's CLIP (trained on data up to 2020) loses $\approx 8\%$ zero-shot accuracy on our curated retrieval task from 2021-2022 compared with more recently trained models in OpenCLIP repository. We then study how to efficiently train models on time-continuous data. We demonstrate that a simple rehearsal-based approach that continues training from the last checkpoint and replays old data reduces compute by $2.5\times$ when compared to the standard practice of retraining from scratch. Code is available at https://github.com/apple/ml-tic-clip.

ICLR Conference 2023 Conference Paper

FastFill: Efficient Compatible Model Update

• Florian Jaeckle
• Fartash Faghri
• Ali Farhadi
• Oncel Tuzel
• Hadi Pouransari

In many retrieval systems the original high dimensional data (e.g., images) is mapped to a lower dimensional feature through a learned embedding model. The task of retrieving the most similar data from a gallery set to a given query data is performed through similarity comparison on features. When the embedding model is updated, it might produce features that are not comparable/compatible with features already in the gallery computed with the old model. Subsequently, all features in the gallery need to be re-computed using the new embedding model -- a computationally expensive process called backfilling. Recently, compatible representation learning methods have been proposed to avoid back-filling. Despite their relative success, there is an inherent trade-off between new model performance and its compatibility with the old model. In this work, we introduce FastFill: a compatible model update process using feature alignment and policy based partial backfilling to promptly elevate retrieval performance. We show that previous backfilling strategies suffer from decreased performance and demonstrate the importance of both the training objective and the ordering in online partial backfilling. We propose a new training method for feature alignment between old and new embedding models using uncertainty estimation. Compared to previous works, we obtain significantly improved backfilling results on a variety of datasets: mAP on ImageNet (+4.4%), Places-365 (+2.7%), and VGG-Face2 (+1.3%). Further, we demonstrate that when updating a biased model with FastFill, the minority subgroup accuracy gap promptly vanishes with a small fraction of partial backfilling.

ICML Conference 2022 Conference Paper

Style Equalization: Unsupervised Learning of Controllable Generative Sequence Models

• Jen-Hao Rick Chang
• Ashish Shrivastava 0001
• Hema Swetha Koppula
• Xiaoshuai Zhang
• Oncel Tuzel

Controllable generative sequence models with the capability to extract and replicate the style of specific examples enable many applications, including narrating audiobooks in different voices, auto-completing and auto-correcting written handwriting, and generating missing training samples for downstream recognition tasks. However, under an unsupervised-style setting, typical training algorithms for controllable sequence generative models suffer from the training-inference mismatch, where the same sample is used as content and style input during training but unpaired samples are given during inference. In this paper, we tackle the training-inference mismatch encountered during unsupervised learning of controllable generative sequence models. The proposed method is simple yet effective, where we use a style transformation module to transfer target style information into an unrelated style input. This method enables training using unpaired content and style samples and thereby mitigate the training-inference mismatch. We apply style equalization to text-to-speech and text-to-handwriting synthesis on three datasets. We conduct thorough evaluation, including both quantitative and qualitative user studies. Our results show that by mitigating the training-inference mismatch with the proposed style equalization, we achieve style replication scores comparable to real data in our user studies.

NeurIPS Conference 2019 Conference Paper

Data Parameters: A New Family of Parameters for Learning a Differentiable Curriculum

• Shreyas Saxena
• Oncel Tuzel
• Dennis DeCoste

Recent works have shown that learning from easier instances first can help deep neural networks (DNNs) generalize better. However, knowing which data to present during different stages of training is a challenging problem. In this work, we address this problem by introducing data parameters. More specifically, we equip each sample and class in a dataset with a learnable parameter (data parameters), which governs their importance in the learning process. During training, at each iteration, as we update the model parameters, we also update the data parameters. These updates are done by gradient descent and do not require hand-crafted rules or design. When applied to image classification task on CIFAR10, CIFAR100, WebVision and ImageNet datasets, and object detection task on KITTI dataset, learning a dynamic curriculum via data parameters leads to consistent gains, without any increase in model complexity or training time. When applied to a noisy dataset, the proposed method learns to learn from clean images and improves over the state-of-the-art methods by 14%. To the best of our knowledge, our work is the first curriculum learning method to show gains on large scale image classification and detection tasks.

ICRA Conference 2019 Conference Paper

MVX-Net: Multimodal VoxelNet for 3D Object Detection

• Vishwanath Sindagi
• Yin Zhou
• Oncel Tuzel

Many recent works on 3D object detection have focused on designing neural network architectures that can consume point cloud data. While these approaches demonstrate encouraging performance, they are typically based on a single modality and are unable to leverage information from other modalities, such as a camera. Although a few approaches fuse data from different modalities, these methods either use a complicated pipeline to process the modalities sequentially, or perform late-fusion and are unable to learn interaction between different modalities at early stages. In this work, we present PointFusion and VoxelFusion: two simple yet effective early-fusion approaches to combine the RGB and point cloud modalities, by leveraging the recently introduced VoxelNet architecture. Evaluation on the KITTI dataset demonstrates significant improvements in performance over approaches which only use point cloud data. Furthermore, the proposed method provides results competitive with the state-of-the-art multimodal algorithms, achieving top-2 ranking in five of the six birds eye view and 3D detection categories on the KITTI benchmark, by using a simple single stage network.

NeurIPS Conference 2016 Conference Paper

Coupled Generative Adversarial Networks

• Ming-Yu Liu
• Oncel Tuzel

We propose the coupled generative adversarial nets (CoGAN) framework for generating pairs of corresponding images in two different domains. The framework consists of a pair of generative adversarial nets, each responsible for generating images in one domain. We show that by enforcing a simple weight-sharing constraint, the CoGAN learns to generate pairs of corresponding images without existence of any pairs of corresponding images in the two domains in the training set. In other words, the CoGAN learns a joint distribution of images in the two domains from images drawn separately from the marginal distributions of the individual domains. This is in contrast to the existing multi-modal generative models, which require corresponding images for training. We apply the CoGAN to several pair image generation tasks. For each task, the CoGAN learns to generate convincing pairs of corresponding images. We further demonstrate the applications of the CoGAN framework for the domain adaptation and cross-domain image generation tasks.

NeurIPS Conference 2014 Conference Paper

Recursive Context Propagation Network for Semantic Scene Labeling

• Abhishek Sharma
• Oncel Tuzel
• Ming-Yu Liu

We propose a deep feed-forward neural network architecture for pixel-wise semantic scene labeling. It uses a novel recursive neural network architecture for context propagation, referred to as rCPN. It first maps the local visual features into a semantic space followed by a bottom-up aggregation of local information into a global representation of the entire image. Then a top-down propagation of the aggregated information takes place that enhances the contextual information of each local feature. Therefore, the information from every location in the image is propagated to every other location. Experimental results on Stanford background and SIFT Flow datasets show that the proposed method outperforms previous approaches. It is also orders of magnitude faster than previous methods and takes only 0. 07 seconds on a GPU for pixel-wise labeling of a 256x256 image starting from raw RGB pixel values, given the super-pixel mask that takes an additional 0. 3 seconds using an off-the-shelf implementation.

ICRA Conference 2012 Conference Paper

Voting-based pose estimation for robotic assembly using a 3D sensor

• Changhyun Choi
• Yuichi Taguchi
• Oncel Tuzel
• Ming-Yu Liu 0001
• Srikumar Ramalingam

We propose a voting-based pose estimation algorithm applicable to 3D sensors, which are fast replacing their 2D counterparts in many robotics, computer vision, and gaming applications. It was recently shown that a pair of oriented 3D points, which are points on the object surface with normals, in a voting framework enables fast and robust pose estimation. Although oriented surface points are discriminative for objects with sufficient curvature changes, they are not compact and discriminative enough for many industrial and real-world objects that are mostly planar. As edges play the key role in 2D registration, depth discontinuities are crucial in 3D. In this paper, we investigate and develop a family of pose estimation algorithms that better exploit this boundary information. In addition to oriented surface points, we use two other primitives: boundary points with directions and boundary line segments. Our experiments show that these carefully chosen primitives encode more information compactly and thereby provide higher accuracy for a wide class of industrial parts and enable faster computation. We demonstrate a practical robotic bin-picking system using the proposed algorithm and a 3D sensor.

UAI Conference 2011 Conference Paper

Compressed Inference for Probabilistic Sequential Models

• Gungor Polatkan
• Oncel Tuzel

ICRA Conference 2011 Conference Paper

Finding a needle in a specular haystack

• Nitesh Shroff
• Yuichi Taguchi
• Oncel Tuzel
• Ashok Veeraraghavan
• Srikumar Ramalingam
• Haruhisa Okuda

Progress in machine vision algorithms has led to widespread adoption of these techniques to automate several industrial assembly tasks. Nevertheless, shiny or specular objects which are common in industrial environments still present a great challenge for vision systems. In this paper, we take a step towards this problem under the context of vision-aided robotic assembly. We show that when the illumination source moves, the specular highlights remain in a region whose radius is inversely proportional to the surface curvature. This allows us to extract regions of the object that have high surface curvature. These points of high curvature can be used as features for specular objects. Further, an inexpensive multi-flash camera (MFC) design can be used to reliably extract these features. We show that one can use multiple views of the object using the MFC in order to triangulate and obtain the 3D location and pose of the shiny objects. Finally, we show a system consisting of a robot arm with an MFC that can perform automated detection and pose estimation of shiny screws within a cluttered bin, achieving position and orientation errors less than 0. 5 mm and 0. 8° respectively.

ICRA Conference 2010 Conference Paper

Pose estimation in heavy clutter using a multi-flash camera

• Ming-Yu Liu 0001
• Oncel Tuzel
• Ashok Veeraraghavan
• Rama Chellappa
• Amit K. Agrawal
• Haruhisa Okuda

We propose a novel solution to object detection, localization and pose estimation with applications in robot vision. The proposed method is especially applicable when the objects of interest may not be richly textured and are immersed in heavy clutter. We show that a multi-flash camera (MFC) provides accurate separation of depth edges and texture edges in such scenes. Then, we reformulate the problem, as one of finding matches between the depth edges obtained in one or more MFC images to the rendered depth edges that are computed offline using 3D CAD model of the objects. In order to facilitate accurate matching of these binary depth edge maps, we introduce a novel cost function that respects both the position and the local orientation of each edge pixel. This cost function is significantly superior to traditional Chamfer cost and leads to accurate matching even in heavily cluttered scenes where traditional methods are unreliable. We present a sub-linear time algorithm to compute the cost function using techniques from 3D distance transforms and integral images. Finally, we also propose a multi-view based pose-refinement algorithm to improve the estimated pose. We implemented the algorithm on an industrial robot arm and obtained location and angular estimation accuracy of the order of 1 mm and 2° respectively for a variety of parts with minimal texture.