Maneesh Singh

Paritosh Mittal, Yen-Chi Cheng, Maneesh Singh et al.

Powerful priors allow us to perform inference with insufficient information. In this paper, we propose an autoregressive prior for 3D shapes to solve multimodal 3D tasks such as shape completion, reconstruction, and generation. We model the distribution over 3D shapes as a non-sequential autoregressive distribution over a discretized, low-dimensional, symbolic grid-like latent representation of 3D shapes. This enables us to represent distributions over 3D shapes conditioned on information from an arbitrary set of spatially anchored query locations and thus perform shape completion in such arbitrary settings (e.g., generating a complete chair given only a view of the back leg). We also show that the learned autoregressive prior can be leveraged for conditional tasks such as single-view reconstruction and language-based generation. This is achieved by learning task-specific naive conditionals which can be approximated by light-weight models trained on minimal paired data. We validate the effectiveness of the proposed method using both quantitative and qualitative evaluation and show that the proposed method outperforms the specialized state-of-the-art methods trained for individual tasks. The project page with code and video visualizations can be found at https://yccyenchicheng.github.io/AutoSDF/.

Sreyan Ghosh, Ashish Seth, and Deepak Mittal et al.

Inspired by the recent progress in self-supervised learning for computer vision, in this paper we introduce DeLoRes, a new general-purpose audio representation learning approach. Our main objective is to make our network learn representations in a resource-constrained setting (both data and compute), that can generalize well across a diverse set of downstream tasks. Inspired from the Barlow Twins objective function, we propose to learn embeddings that are invariant to distortions of an input audio sample, while making sure that they contain non-redundant information about the sample. To achieve this, we measure the cross-correlation matrix between the outputs of two identical networks fed with distorted versions of an audio segment sampled from an audio file and make it as close to the identity matrix as possible. We use a combination of a small subset of the large-scale AudioSet dataset and FSD50K for self-supervised learning and are able to learn with less than half the parameters compared to state-of-the-art algorithms. For evaluation, we transfer these learned representations to 9 downstream classification tasks, including speech, music, and animal sounds, and show competitive results under different evaluation setups. In addition to being simple and intuitive, our pre-training algorithm is amenable to compute through its inherent nature of construction and does not require careful implementation details to avoid trivial or degenerate solutions. Furthermore, we conduct ablation studies on our results and make all our code and pre-trained models publicly available https://github.com/Speech-Lab-IITM/DeLoRes.

Debottam Dutta, Debarpan Bhattacharya, Sriram Ganapathy et al. · deepmind

In this paper, we describe an approach for representation learning of audio signals for the task of COVID-19 detection. The raw audio samples are processed with a bank of 1-D convolutional filters that are parameterized as cosine modulated Gaussian functions. The choice of these kernels allows the interpretation of the filterbanks as smooth band-pass filters. The filtered outputs are pooled, log-compressed and used in a self-attention based relevance weighting mechanism. The relevance weighting emphasizes the key regions of the time-frequency decomposition that are important for the downstream task. The subsequent layers of the model consist of a recurrent architecture and the models are trained for a COVID-19 detection task. In our experiments on the Coswara data set, we show that the proposed model achieves significant performance improvements over the baseline system as well as other representation learning approaches. Further, the approach proposed is shown to be uniformly applicable for speech and breathing signals and for transfer learning from a larger data set.

Deepak Mittal, Amir H. Poorjam, Debottam Dutta et al. · deepmind

This report describes the system used for detecting COVID-19 positives using three different acoustic modalities, namely speech, breathing, and cough in the second DiCOVA challenge. The proposed system is based on the combination of 4 different approaches, each focusing more on one aspect of the problem, and reaches the blind test AUCs of 86.41, 77.60, and 84.55, in the breathing, cough, and speech tracks, respectively, and the AUC of 85.37 in the fusion of these three tracks.

Susmit Agrawal, Prabhat Kumar, Siddharth Seth et al.

Recent algorithms for image manipulation detection almost exclusively use deep network models. These approaches require either dense pixelwise groundtruth masks, camera ids, or image metadata to train the networks. On one hand, constructing a training set to represent the countless tampering possibilities is impractical. On the other hand, social media platforms or commercial applications are often constrained to remove camera ids as well as metadata from images. A self-supervised algorithm for training manipulation detection models without dense groundtruth or camera/image metadata would be extremely useful for many forensics applications. In this paper, we propose self-supervised approach for training splicing detection/localization models from frequency transforms of images. To identify the spliced regions, our deep network learns a representation to capture an image specific signature by enforcing (image) self consistency . We experimentally demonstrate that our proposed model can yield similar or better performances of multiple existing methods on standard datasets without relying on labels or metadata.

Tejan Karmali, Rishubh Parihar, Susmit Agrawal et al.

Progress in GANs has enabled the generation of high-resolution photorealistic images of astonishing quality. StyleGANs allow for compelling attribute modification on such images via mathematical operations on the latent style vectors in the W/W+ space that effectively modulate the rich hierarchical representations of the generator. Such operations have recently been generalized beyond mere attribute swapping in the original StyleGAN paper to include interpolations. In spite of many significant improvements in StyleGANs, they are still seen to generate unnatural images. The quality of the generated images is predicated on two assumptions; (a) The richness of the hierarchical representations learnt by the generator, and, (b) The linearity and smoothness of the style spaces. In this work, we propose a Hierarchical Semantic Regularizer (HSR) which aligns the hierarchical representations learnt by the generator to corresponding powerful features learnt by pretrained networks on large amounts of data. HSR is shown to not only improve generator representations but also the linearity and smoothness of the latent style spaces, leading to the generation of more natural-looking style-edited images. To demonstrate improved linearity, we propose a novel metric - Attribute Linearity Score (ALS). A significant reduction in the generation of unnatural images is corroborated by improvement in the Perceptual Path Length (PPL) metric by 16.19% averaged across different standard datasets while simultaneously improving the linearity of attribute-change in the attribute editing tasks.

Radu Balan, Naveed Haghani, Maneesh Singh

This paper presents primarily two Euclidean embeddings of the quotient space generated by matrices that are identified modulo arbitrary row permutations. The original application is in deep learning on graphs where the learning task is invariant to node relabeling. Two embedding schemes are introduced, one based on sorting and the other based on algebras of multivariate polynomials. While both embeddings exhibit a computational complexity exponential in problem size, the sorting based embedding is globally bi-Lipschitz and admits a low dimensional target space. Additionally, an almost everywhere injective scheme can be implemented with minimal redundancy and low computational cost. In turn, this proves that almost any classifier can be implemented with an arbitrary small loss of performance. Numerical experiments are carried out on two data sets, a chemical compound data set (QM9) and a proteins data set (PROTEINS).

Naveen Venkat, Mayank Agarwal, Maneesh Singh et al.

We tackle the task of synthesizing novel views of an object given a few input images and associated camera viewpoints. Our work is inspired by recent 'geometry-free' approaches where multi-view images are encoded as a (global) set-latent representation, which is then used to predict the color for arbitrary query rays. While this representation yields (coarsely) accurate images corresponding to novel viewpoints, the lack of geometric reasoning limits the quality of these outputs. To overcome this limitation, we propose 'Geometry-biased Transformers' (GBTs) that incorporate geometric inductive biases in the set-latent representation-based inference to encourage multi-view geometric consistency. We induce the geometric bias by augmenting the dot-product attention mechanism to also incorporate 3D distances between rays associated with tokens as a learnable bias. We find that this, along with camera-aware embeddings as input, allows our models to generate significantly more accurate outputs. We validate our approach on the real-world CO3D dataset, where we train our system over 10 categories and evaluate its view-synthesis ability for novel objects as well as unseen categories. We empirically validate the benefits of the proposed geometric biases and show that our approach significantly improves over prior works.

Chieh Hubert Lin, Hsin-Ying Lee, Hung-Yu Tseng et al.

Recent studies show that paddings in convolutional neural networks encode absolute position information which can negatively affect the model performance for certain tasks. However, existing metrics for quantifying the strength of positional information remain unreliable and frequently lead to erroneous results. To address this issue, we propose novel metrics for measuring (and visualizing) the encoded positional information. We formally define the encoded information as PPP (Position-information Pattern from Padding) and conduct a series of experiments to study its properties as well as its formation. The proposed metrics measure the presence of positional information more reliably than the existing metrics based on PosENet and a test in F-Conv. We also demonstrate that for any extant (and proposed) padding schemes, PPP is primarily a learning artifact and is less dependent on the characteristics of the underlying padding schemes.

Subba Reddy Oota, Khushbu Pahwa, Prachi Jindal et al.

Recent voxel-wise multimodal brain encoding studies have shown that multimodal large language models (MLLMs) exhibit a higher degree of brain alignment compared to unimodal models in both unimodal and multimodal stimulus settings. More recently, instruction-tuned multimodal models have shown to generate task-specific representations that align strongly with brain activity. However, prior work evaluating the brain alignment of MLLMs has primarily focused on unimodal settings or relied on non-instruction-tuned multimodal models for multimodal stimuli. To address this gap, we investigated brain alignment, that is, measuring the degree of predictivity of neural activity recorded while participants were watching naturalistic movies (video along with audio) with representations derived from MLLMs. We utilized instruction-specific embeddings from six video and two audio instruction-tuned MLLMs. Experiments with 13 video task-specific instructions show that instruction-tuned video MLLMs significantly outperform non-instruction-tuned multimodal (by 15%) and unimodal models (by 20%). Our evaluation of MLLMs for both video and audio tasks using language-guided instructions shows clear disentanglement in task-specific representations from MLLMs, leading to precise differentiation of multimodal functional processing in the brain. We also find that MLLM layers align hierarchically with the brain, with early sensory areas showing strong alignment with early layers, while higher-level visual and language regions align more with middle to late layers. These findings provide clear evidence for the role of task-specific instructions in improving the alignment between brain activity and MLLMs, and open new avenues for mapping joint information processing in both the systems. We make the code publicly available [https://github.com/subbareddy248/mllm_videos].

Ameya Prabhu, Charles Dognin, Maneesh Singh

The exploding cost and time needed for data labeling and model training are bottlenecks for training DNN models on large datasets. Identifying smaller representative data samples with strategies like active learning can help mitigate such bottlenecks. Previous works on active learning in NLP identify the problem of sampling bias in the samples acquired by uncertainty-based querying and develop costly approaches to address it. Using a large empirical study, we demonstrate that active set selection using the posterior entropy of deep models like FastText.zip (FTZ) is robust to sampling biases and to various algorithmic choices (query size and strategies) unlike that suggested by traditional literature. We also show that FTZ based query strategy produces sample sets similar to those from more sophisticated approaches (e.g ensemble networks). Finally, we show the effectiveness of the selected samples by creating tiny high-quality datasets, and utilizing them for fast and cheap training of large models. Based on the above, we propose a simple baseline for deep active text classification that outperforms the state-of-the-art. We expect the presented work to be useful and informative for dataset compression and for problems involving active, semi-supervised or online learning scenarios. Code and models are available at: https://github.com/drimpossible/Sampling-Bias-Active-Learning

Vishnu Suresh Lokhande, Shaofei Wang, Maneesh Singh et al.

In this paper, we introduce a new optimization approach to Entity Resolution. Traditional approaches tackle entity resolution with hierarchical clustering, which does not benefit from a formal optimization formulation. In contrast, we model entity resolution as correlation-clustering, which we treat as a weighted set-packing problem and write as an integer linear program (ILP). In this case sources in the input data correspond to elements and entities in output data correspond to sets/clusters. We tackle optimization of weighted set packing by relaxing integrality in our ILP formulation. The set of potential sets/clusters can not be explicitly enumerated, thus motivating optimization via column generation. In addition to the novel formulation, we also introduce new dual optimal inequalities (DOI), that we call flexible dual optimal inequalities, which tightly lower-bound dual variables during optimization and accelerate column generation. We apply our formulation to entity resolution (also called de-duplication of records), and achieve state-of-the-art accuracy on two popular benchmark datasets. The project page is available at the following url, https://github.com/lokhande-vishnu/EntityResolution

Yash Malviya, Karan Dhingra, Maneesh Singh

Regulatory documents are rich in nuanced terminology and specialized semantics. FRAG systems: Frozen retrieval-augmented generators utilizing pre-trained (or, frozen) components face consequent challenges with both retriever and answering performance. We present a system that adapts the retriever performance to the target domain using a multi-stage tuning (MST) strategy. Our retrieval approach, called MST-R (a) first fine-tunes encoders used in vector stores using hard negative mining, (b) then uses a hybrid retriever, combining sparse and dense retrievers using reciprocal rank fusion, and then (c) adapts the cross-attention encoder by fine-tuning only the top-k retrieved results. We benchmark the system performance on the dataset released for the RIRAG challenge (as part of the RegNLP workshop at COLING 2025). We achieve significant performance gains obtaining a top rank on the RegNLP challenge leaderboard. We also show that a trivial answering approach games the RePASs metric outscoring all baselines and a pre-trained Llama model. Analyzing this anomaly, we present important takeaways for future research.

Soumya Banerjee, Vinay Kumar Verma, Toufiq Parag et al.

A wide variety of methods have been developed to enable lifelong learning in conventional deep neural networks. However, to succeed, these methods require a `batch' of samples to be available and visited multiple times during training. While this works well in a static setting, these methods continue to suffer in a more realistic situation where data arrives in \emph{online streaming manner}. We empirically demonstrate that the performance of current approaches degrades if the input is obtained as a stream of data with the following restrictions: $(i)$ each instance comes one at a time and can be seen only once, and $(ii)$ the input data violates the i.i.d assumption, i.e., there can be a class-based correlation. We propose a novel approach (CIOSL) for the class-incremental learning in an \emph{online streaming setting} to address these challenges. The proposed approach leverages implicit and explicit dual weight regularization and experience replay. The implicit regularization is leveraged via the knowledge distillation, while the explicit regularization incorporates a novel approach for parameter regularization by learning the joint distribution of the buffer replay and the current sample. Also, we propose an efficient online memory replay and replacement buffer strategy that significantly boosts the model's performance. Extensive experiments and ablation on challenging datasets show the efficacy of the proposed method.

Arjit Jain, Pranay Reddy Samala, Deepak Mittal et al.

Time masking has become a de facto augmentation technique for speech and audio tasks, including automatic speech recognition (ASR) and audio classification, most notably as a part of SpecAugment. In this work, we propose SpliceOut, a simple modification to time masking which makes it computationally more efficient. SpliceOut performs comparably to (and sometimes outperforms) SpecAugment on a wide variety of speech and audio tasks, including ASR for seven different languages using varying amounts of training data, as well as on speech translation, sound and music classification, thus establishing itself as a broadly applicable audio augmentation method. SpliceOut also provides additional gains when used in conjunction with other augmentation techniques. Apart from the fully-supervised setting, we also demonstrate that SpliceOut can complement unsupervised representation learning with performance gains in the semi-supervised and self-supervised settings.

Vivek Gupta, Riyaz A. Bhat, Atreya Ghosal et al.

Neural models command state-of-the-art performance across NLP tasks, including ones involving "reasoning". Models claiming to reason about the evidence presented to them should attend to the correct parts of the input avoiding spurious patterns therein, be self-consistent in their predictions across inputs, and be immune to biases derived from their pre-training in a nuanced, context-sensitive fashion. {\em Do the prevalent *BERT-family of models do so?} In this paper, we study this question using the problem of reasoning on tabular data. Tabular inputs are especially well-suited for the study -- they admit systematic probes targeting the properties listed above. Our experiments demonstrate that a RoBERTa-based model, representative of the current state-of-the-art, fails at reasoning on the following counts: it (a) ignores relevant parts of the evidence, (b) is over-sensitive to annotation artifacts, and (c) relies on the knowledge encoded in the pre-trained language model rather than the evidence presented in its tabular inputs. Finally, through inoculation experiments, we show that fine-tuning the model on perturbed data does not help it overcome the above challenges.

Rahul Venkatesh, Tejan Karmali, Sarthak Sharma et al.

Deep neural representations of 3D shapes as implicit functions have been shown to produce high fidelity models surpassing the resolution-memory trade-off faced by the explicit representations using meshes and point clouds. However, most such approaches focus on representing closed shapes. Unsigned distance function (UDF) based approaches have been proposed recently as a promising alternative to represent both open and closed shapes. However, since the gradients of UDFs vanish on the surface, it is challenging to estimate local (differential) geometric properties like the normals and tangent planes which are needed for many downstream applications in vision and graphics. There are additional challenges in computing these properties efficiently with a low-memory footprint. This paper presents a novel approach that models such surfaces using a new class of implicit representations called the closest surface-point (CSP) representation. We show that CSP allows us to represent complex surfaces of any topology (open or closed) with high fidelity. It also allows for accurate and efficient computation of local geometric properties. We further demonstrate that it leads to efficient implementation of downstream algorithms like sphere-tracing for rendering the 3D surface as well as to create explicit mesh-based representations. Extensive experimental evaluation on the ShapeNet dataset validate the above contributions with results surpassing the state-of-the-art.

Hsin-Ping Huang, Hung-Yu Tseng, Saurabh Saini et al.

We tackle a 3D scene stylization problem - generating stylized images of a scene from arbitrary novel views given a set of images of the same scene and a reference image of the desired style as inputs. Direct solution of combining novel view synthesis and stylization approaches lead to results that are blurry or not consistent across different views. We propose a point cloud-based method for consistent 3D scene stylization. First, we construct the point cloud by back-projecting the image features to the 3D space. Second, we develop point cloud aggregation modules to gather the style information of the 3D scene, and then modulate the features in the point cloud with a linear transformation matrix. Finally, we project the transformed features to 2D space to obtain the novel views. Experimental results on two diverse datasets of real-world scenes validate that our method generates consistent stylized novel view synthesis results against other alternative approaches.

Yichen Jiang, Shikha Bordia, Zheng Zhong et al.

We introduce HoVer (HOppy VERification), a dataset for many-hop evidence extraction and fact verification. It challenges models to extract facts from several Wikipedia articles that are relevant to a claim and classify whether the claim is Supported or Not-Supported by the facts. In HoVer, the claims require evidence to be extracted from as many as four English Wikipedia articles and embody reasoning graphs of diverse shapes. Moreover, most of the 3/4-hop claims are written in multiple sentences, which adds to the complexity of understanding long-range dependency relations such as coreference. We show that the performance of an existing state-of-the-art semantic-matching model degrades significantly on our dataset as the number of reasoning hops increases, hence demonstrating the necessity of many-hop reasoning to achieve strong results. We hope that the introduction of this challenging dataset and the accompanying evaluation task will encourage research in many-hop fact retrieval and information verification. We make the HoVer dataset publicly available at https://hover-nlp.github.io

Rahul Venkatesh, Sarthak Sharma, Aurobrata Ghosh et al.

High fidelity representation of shapes with arbitrary topology is an important problem for a variety of vision and graphics applications. Owing to their limited resolution, classical discrete shape representations using point clouds, voxels and meshes produce low quality results when used in these applications. Several implicit 3D shape representation approaches using deep neural networks have been proposed leading to significant improvements in both quality of representations as well as the impact on downstream applications. However, these methods can only be used to represent topologically closed shapes which greatly limits the class of shapes that they can represent. As a consequence, they also often require clean, watertight meshes for training. In this work, we propose DUDE - a Deep Unsigned Distance Embedding method which alleviates both of these shortcomings. DUDE is a disentangled shape representation that utilizes an unsigned distance field (uDF) to represent proximity to a surface, and a normal vector field (nVF) to represent surface orientation. We show that a combination of these two (uDF+nVF) can be used to learn high fidelity representations for arbitrary open/closed shapes. As opposed to prior work such as DeepSDF, our shape representations can be directly learnt from noisy triangle soups, and do not need watertight meshes. Additionally, we propose novel algorithms for extracting and rendering iso-surfaces from the learnt representations. We validate DUDE on benchmark 3D datasets and demonstrate that it produces significant improvements over the state of the art.

VSR Veeravasarapu, Abhishek Goel, Deepak Mittal et al.

Contour shape alignment is a fundamental but challenging problem in computer vision, especially when the observations are partial, noisy, and largely misaligned. Recent ConvNet-based architectures that were proposed to align image structures tend to fail with contour representation of shapes, mostly due to the use of proximity-insensitive pixel-wise similarity measures as loss functions in their training processes. This work presents a novel ConvNet, "ProAlignNet" that accounts for large scale misalignments and complex transformations between the contour shapes. It infers the warp parameters in a multi-scale fashion with progressively increasing complex transformations over increasing scales. It learns --without supervision-- to align contours, agnostic to noise and missing parts, by training with a novel loss function which is derived an upperbound of a proximity-sensitive and local shape-dependent similarity metric that uses classical Morphological Chamfer Distance Transform. We evaluate the reliability of these proposals on a simulated MNIST noisy contours dataset via some basic sanity check experiments. Next, we demonstrate the effectiveness of the proposed models in two real-world applications of (i) aligning geo-parcel data to aerial image maps and (ii) refining coarsely annotated segmentation labels. In both applications, the proposed models consistently perform superior to state-of-the-art methods.

Shaofei Wang, Vishnu Lokhande, Maneesh Singh et al.

Modern computer vision (CV) is often based on convolutional neural networks (CNNs) that excel at hierarchical feature extraction. The previous generation of CV approaches was often based on conditional random fields (CRFs) that excel at modeling flexible higher order interactions. As their benefits are complementary they are often combined. However, these approaches generally use mean-field approximations and thus, arguably, did not directly optimize the real problem. Here we revisit dual-decomposition-based approaches to CRF optimization, an alternative to the mean-field approximation. These algorithms can efficiently and exactly solve sub-problems and directly optimize a convex upper bound of the real problem, providing optimality certificates on the way. Our approach uses a novel fixed-point iteration algorithm which enjoys dual-monotonicity, dual-differentiability and high parallelism. The whole system, CRF and CNN can thus be efficiently trained using back-propagation. We demonstrate the effectiveness of our system on semantic image segmentation, showing consistent improvement over baseline models.

Prashnna K Gyawali, Rudra Saha, Linwei Wang et al.

Variational Autoencoders (VAE) are probabilistic deep generative models underpinned by elegant theory, stable training processes, and meaningful manifold representations. However, they produce blurry images due to a lack of explicit emphasis over high-frequency textural details of the images, and the difficulty to directly model the complex joint probability distribution over the high-dimensional image space. In this work, we approach these two challenges with a novel wavelet space VAE that uses the decoder to model the images in the wavelet coefficient space. This enables the VAE to emphasize over high-frequency components within an image obtained via wavelet decomposition. Additionally, by decomposing the complex function of generating high-dimensional images into inverse wavelet transformation and generation of wavelet coefficients, the latter becomes simpler to model by the VAE. We empirically validate that deep generative models operating in the wavelet space can generate images of higher quality than the image (RGB) space counterparts. Quantitatively, on benchmark natural image datasets, we achieve consistently better FID scores than VAE based architectures and competitive FID scores with a variety of GAN models for the same architectural and experimental setup. Furthermore, the proposed wavelet-based generative model retains desirable attributes like disentangled and informative latent representation without losing the quality in the generated samples.

Han-Kai Hsu, Chun-Han Yao, Yi-Hsuan Tsai et al.

Recent deep learning methods for object detection rely on a large amount of bounding box annotations. Collecting these annotations is laborious and costly, yet supervised models do not generalize well when testing on images from a different distribution. Domain adaptation provides a solution by adapting existing labels to the target testing data. However, a large gap between domains could make adaptation a challenging task, which leads to unstable training processes and sub-optimal results. In this paper, we propose to bridge the domain gap with an intermediate domain and progressively solve easier adaptation subtasks. This intermediate domain is constructed by translating the source images to mimic the ones in the target domain. To tackle the domain-shift problem, we adopt adversarial learning to align distributions at the feature level. In addition, a weighted task loss is applied to deal with unbalanced image quality in the intermediate domain. Experimental results show that our method performs favorably against the state-of-the-art method in terms of the performance on the target domain.

Yufei Ye, Maneesh Singh, Abhinav Gupta et al.

We present an approach for pixel-level future prediction given an input image of a scene. We observe that a scene is comprised of distinct entities that undergo motion and present an approach that operationalizes this insight. We implicitly predict future states of independent entities while reasoning about their interactions, and compose future video frames using these predicted states. We overcome the inherent multi-modality of the task using a global trajectory-level latent random variable, and show that this allows us to sample diverse and plausible futures. We empirically validate our approach against alternate representations and ways of incorporating multi-modality. We examine two datasets, one comprising of stacked objects that may fall, and the other containing videos of humans performing activities in a gym, and show that our approach allows realistic stochastic video prediction across these diverse settings. See https://judyye.github.io/CVP/ for video predictions.

Aurobrata Ghosh, Zheng Zhong, Steve Cruz et al.

We consider an information theoretic approach to address the problem of identifying fake digital images. We propose an innovative method to formulate the issue of localizing manipulated regions in an image as a deep representation learning problem using the Information Bottleneck (IB), which has recently gained popularity as a framework for interpreting deep neural networks. Tampered images pose a serious predicament since digitized media is a ubiquitous part of our lives. These are facilitated by the easy availability of image editing software and aggravated by recent advances in deep generative models such as GANs. We propose InfoPrint, a computationally efficient solution to the IB formulation using approximate variational inference and compare it to a numerical solution that is computationally expensive. Testing on a number of standard datasets, we demonstrate that InfoPrint outperforms the state-of-the-art and the numerical solution. Additionally, it also has the ability to detect alterations made by inpainting GANs.

Aurobrata Ghosh, Zheng Zhong, Terrance E Boult et al.

Detection and localization of image manipulations like splices are gaining in importance with the easy accessibility of image editing softwares. While detection generates a verdict for an image it provides no insight into the manipulation. Localization helps explain a positive detection by identifying the pixels of the image which have been tampered. We propose a deep learning based method for splice localization without prior knowledge of a test image's camera-model. It comprises a novel approach for learning rich filters and for suppressing image-edges. Additionally, we train our model on a surrogate task of camera model identification, which allows us to leverage large and widely available, unmanipulated, camera-tagged image databases. During inference, we assume that the spliced and host regions come from different camera-models and we segment these regions using a Gaussian-mixture model. Experiments on three test databases demonstrate results on par with and above the state-of-the-art and a good generalization ability to unknown datasets.

Hsin-Ying Lee, Hung-Yu Tseng, Qi Mao et al.

Image-to-image translation aims to learn the mapping between two visual domains. There are two main challenges for this task: 1) lack of aligned training pairs and 2) multiple possible outputs from a single input image. In this work, we present an approach based on disentangled representation for generating diverse outputs without paired training images. To synthesize diverse outputs, we propose to embed images onto two spaces: a domain-invariant content space capturing shared information across domains and a domain-specific attribute space. Our model takes the encoded content features extracted from a given input and attribute vectors sampled from the attribute space to synthesize diverse outputs at test time. To handle unpaired training data, we introduce a cross-cycle consistency loss based on disentangled representations. Qualitative results show that our model can generate diverse and realistic images on a wide range of tasks without paired training data. For quantitative evaluations, we measure realism with user study and Fréchet inception distance, and measure diversity with the perceptual distance metric, Jensen-Shannon divergence, and number of statistically-different bins.

Margret Keuper, Jovita Lukasik, Maneesh Singh et al.

We tackle the problem of graph partitioning for image segmentation using correlation clustering (CC), which we treat as an integer linear program (ILP). We reformulate optimization in the ILP so as to admit efficient optimization via Benders decomposition, a classic technique from operations research. Our Benders decomposition formulation has many subproblems, each associated with a node in the CC instance's graph, which are solved in parallel. Each Benders subproblem enforces the cycle inequalities corresponding to the negative weight edges attached to its corresponding node in the CC instance. We generate Magnanti-Wong Benders rows in addition to standard Benders rows, to accelerate optimization. Our Benders decomposition approach provides a promising new avenue to accelerate optimization for CC, and allows for massive parallelization.

Dongmian Zou, Radu Balan, Maneesh Singh

Many convolutional neural networks (CNNs) have a feed-forward structure. In this paper, a linear program that estimates the Lipschitz bound of such CNNs is proposed. Several CNNs, including the scattering networks, the AlexNet and the GoogleNet, are studied numerically and compared to the theoretical bounds. Next, concentration inequalities of the output distribution to a stationary random input signal expressed in terms of the Lipschitz bound are established. The Lipschitz bound is further used to establish a nonlinear discriminant analysis designed to measure the separation between features of different classes.

Mahyar Khayatkhoei, Ahmed Elgammal, Maneesh Singh

Natural images may lie on a union of disjoint manifolds rather than one globally connected manifold, and this can cause several difficulties for the training of common Generative Adversarial Networks (GANs). In this work, we first show that single generator GANs are unable to correctly model a distribution supported on a disconnected manifold, and investigate how sample quality, mode dropping and local convergence are affected by this. Next, we show how using a collection of generators can address this problem, providing new insights into the success of such multi-generator GANs. Finally, we explain the serious issues caused by considering a fixed prior over the collection of generators and propose a novel approach for learning the prior and inferring the necessary number of generators without any supervision. Our proposed modifications can be applied on top of any other GAN model to enable learning of distributions supported on disconnected manifolds. We conduct several experiments to illustrate the aforementioned shortcoming of GANs, its consequences in practice, and the effectiveness of our proposed modifications in alleviating these issues.

Aditay Tripathi, Aanchan Mohan, Saket Anand et al.

Recent advances in neural network based acoustic modelling have shown significant improvements in automatic speech recognition (ASR) performance. In order for acoustic models to be able to handle large acoustic variability, large amounts of labeled data is necessary, which are often expensive to obtain. This paper explores the application of adversarial training to learn features from raw speech that are invariant to acoustic variability. This acoustic variability is referred to as a domain shift in this paper. The experimental study presented in this paper leverages the architecture of Domain Adversarial Neural Networks (DANNs) [1] which uses data from two different domains. The DANN is a Y-shaped network that consists of a multi-layer CNN feature extractor module that is common to a label (senone) classifier and a so-called domain classifier. The utility of DANNs is evaluated on multiple datasets with domain shifts caused due to differences in gender and speaker accents. Promising empirical results indicate the strength of adversarial training for unsupervised domain adaptation in ASR, thereby emphasizing the ability of DANNs to learn domain invariant features from raw speech.

Ananya Harsh Jha, Saket Anand, Maneesh Singh et al.

Generative models that learn disentangled representations for different factors of variation in an image can be very useful for targeted data augmentation. By sampling from the disentangled latent subspace of interest, we can efficiently generate new data necessary for a particular task. Learning disentangled representations is a challenging problem, especially when certain factors of variation are difficult to label. In this paper, we introduce a novel architecture that disentangles the latent space into two complementary subspaces by using only weak supervision in form of pairwise similarity labels. Inspired by the recent success of cycle-consistent adversarial architectures, we use cycle-consistency in a variational auto-encoder framework. Our non-adversarial approach is in contrast with the recent works that combine adversarial training with auto-encoders to disentangle representations. We show compelling results of disentangled latent subspaces on three datasets and compare with recent works that leverage adversarial training.

Hsin-Ying Lee, Jia-Bin Huang, Maneesh Singh et al.

We present an unsupervised representation learning approach using videos without semantic labels. We leverage the temporal coherence as a supervisory signal by formulating representation learning as a sequence sorting task. We take temporally shuffled frames (i.e., in non-chronological order) as inputs and train a convolutional neural network to sort the shuffled sequences. Similar to comparison-based sorting algorithms, we propose to extract features from all frame pairs and aggregate them to predict the correct order. As sorting shuffled image sequence requires an understanding of the statistical temporal structure of images, training with such a proxy task allows us to learn rich and generalizable visual representation. We validate the effectiveness of the learned representation using our method as pre-training on high-level recognition problems. The experimental results show that our method compares favorably against state-of-the-art methods on action recognition, image classification and object detection tasks.

Radu Balan, Maneesh Singh, Dongmian Zou

In this paper we discuss the stability properties of convolutional neural networks. Convolutional neural networks are widely used in machine learning. In classification they are mainly used as feature extractors. Ideally, we expect similar features when the inputs are from the same class. That is, we hope to see a small change in the feature vector with respect to a deformation on the input signal. This can be established mathematically, and the key step is to derive the Lipschitz properties. Further, we establish that the stability results can be extended for more general networks. We give a formula for computing the Lipschitz bound, and compare it with other methods to show it is closer to the optimal value.