Shivam Chandhok

Tushar Choudhary, Vikrant Dewangan, Shivam Chandhok et al. · mit

Talk2BEV is a large vision-language model (LVLM) interface for bird's-eye view (BEV) maps in autonomous driving contexts. While existing perception systems for autonomous driving scenarios have largely focused on a pre-defined (closed) set of object categories and driving scenarios, Talk2BEV blends recent advances in general-purpose language and vision models with BEV-structured map representations, eliminating the need for task-specific models. This enables a single system to cater to a variety of autonomous driving tasks encompassing visual and spatial reasoning, predicting the intents of traffic actors, and decision-making based on visual cues. We extensively evaluate Talk2BEV on a large number of scene understanding tasks that rely on both the ability to interpret free-form natural language queries, and in grounding these queries to the visual context embedded into the language-enhanced BEV map. To enable further research in LVLMs for autonomous driving scenarios, we develop and release Talk2BEV-Bench, a benchmark encompassing 1000 human-annotated BEV scenarios, with more than 20,000 questions and ground-truth responses from the NuScenes dataset.

Shivam Chandhok, Pranav Tandon

Recent advances in vision-language foundational models have enabled development of systems that can perform visual understanding and reasoning tasks. However, it is unclear if these models are robust to distribution shifts, and how their performance and generalization capabilities vary under changes in data distribution. In this project we strive to answer the question "Are vision-language foundational models robust to distribution shifts like human perception?" Specifically, we consider a diverse range of vision-language models and compare how the performance of these systems is affected by corruption based distribution shifts (such as \textit{motion blur, fog, snow, gaussian noise}) commonly found in practical real-world scenarios. We analyse the generalization capabilities qualitatively and quantitatively on zero-shot image classification task under aforementioned distribution shifts. Our code will be avaible at \url{https://github.com/shivam-chandhok/CPSC-540-Project}

Puneet Mangla, Shivam Chandhok, Milan Aggarwal et al.

For models to generalize under unseen domains (a.k.a domain generalization), it is crucial to learn feature representations that are domain-agnostic and capture the underlying semantics that makes up an object category. Recent advances towards weakly supervised vision-language models that learn holistic representations from cheap weakly supervised noisy text annotations have shown their ability on semantic understanding by capturing object characteristics that generalize under different domains. However, when multiple source domains are involved, the cost of curating textual annotations for every image in the dataset can blow up several times, depending on their number. This makes the process tedious and infeasible, hindering us from directly using these supervised vision-language approaches to achieve the best generalization on an unseen domain. Motivated from this, we study how multimodal information from existing pre-trained multimodal networks can be leveraged in an "intrinsic" way to make systems generalize under unseen domains. To this end, we propose IntriNsic multimodality for DomaIn GeneralizatiOn (INDIGO), a simple and elegant way of leveraging the intrinsic modality present in these pre-trained multimodal networks along with the visual modality to enhance generalization to unseen domains at test-time. We experiment on several Domain Generalization settings (ClosedDG, OpenDG, and Limited sources) and show state-of-the-art generalization performance on unseen domains. Further, we provide a thorough analysis to develop a holistic understanding of INDIGO.

Shivam Chandhok

Building models that can understand and reason about 3D scenes is difficult owing to the lack of data sources for 3D supervised training and large-scale training regimes. In this work we ask - How can the knowledge in a pre-trained language model be leveraged for 3D scene understanding without any 3D pre-training. The aim of this work is to establish whether pre-trained LLMs possess priors/knowledge required for reasoning in 3D space and how can we prompt them such that they can be used for general purpose spatial reasoning and object understanding in 3D. To this end, we present SceneGPT, an LLM based scene understanding system which can perform 3D spatial reasoning without training or explicit 3D supervision. The key components of our framework are - 1) a 3D scene graph, that serves as scene representation, encoding the objects in the scene and their spatial relationships 2) a pre-trained LLM that can be adapted with in context learning for 3D spatial reasoning. We evaluate our framework qualitatively on object and scene understanding tasks including object semantics, physical properties and affordances (object-level) and spatial understanding (scene-level).

Shivam Chandhok, Wan-Cyuan Fan, Leonid Sigal

Vision-Language Models (VLMs) have emerged as general purpose tools for addressing a variety of complex computer vision problems. Such models have been shown to be highly capable, but, at the same time, also lacking some basic visual understanding skills. In this paper, we set out to understand the limitations of SoTA VLMs on fundamental visual tasks: object classification, understanding spatial arrangement, and ability to delineate individual object instances (through counting), by constructing a series of tests that probe which components of design, specifically, maybe lacking. Importantly, we go significantly beyond the current benchmarks, that simply measure final performance of VLM, by also comparing and contrasting it to performance of probes trained directly on features obtained from visual encoder (image embeddings), as well as intermediate vision-language projection used to bridge image-encoder and LLM-decoder ouput in many SoTA models (e.g., LLaVA, BLIP, InstructBLIP). In doing so, we uncover nascent shortcomings in VLMs response and make a number of important observations which could help train and develop more effective VLM models in future.

Shivam Chandhok, Qian Yang, Oscar Manas et al.

Instruction tuning has been central to the success of recent vision-language models (VLMs), but it remains expensive-requiring large-scale datasets, high-quality annotations, and large compute budgets. We propose PRioritized cOncept learninG via Relative Error-driven Sample Selection (PROGRESS), a data- and compute-efficient framework that enables VLMs to dynamically select what to learn next based on their evolving needs during training. At each stage, the model tracks its learning progress across skills and selects the most informative samples-those it has not already mastered and that are not too difficult to learn at the current stage of training. This strategy effectively controls skill acquisition and the order in which skills are learned. Specifically, we sample from skills showing the highest learning progress, prioritizing those with the most rapid improvement. Unlike prior methods, PROGRESS requires no upfront answer annotations, queries answers only on a need basis, avoids reliance on additional supervision from auxiliary VLMs, and does not require compute-heavy gradient computations for data selection. Experiments across multiple instruction-tuning datasets of varying scales demonstrate that PROGRESS consistently outperforms state-of-the-art baselines with much less data and supervision. Additionally, we show strong cross-architecture generalization and transferability to larger models, validating PROGRESS as a scalable solution for efficient learning.

Shivam Chandhok, Wan-Cyuan Fan, Vered Shwartz et al.

Vision-language Models (VLMs) have emerged as general-purpose tools for addressing a variety of complex computer vision problems. Such models have been shown to be highly capable, but, at the same time, lacking some basic visual understanding skills. In this paper, we set out to understand the limitations of SoTA VLMs on fundamental visual tasks by constructing a series of tests that probe which components of design, specifically, may be lacking. Importantly, we go significantly beyond the current benchmarks, which simply measure the final performance of VLM response, by also comparing and contrasting it to the performance of probes trained directly on features obtained from the visual encoder, intermediate vision-language projection and LLM-decoder output. In doing so, we uncover shortcomings in VLMs and make a number of important observations about their capabilities, robustness and how they process visual information. We hope our insights will guide progress in further improving VLMs.

Shih-Han Chou, Shivam Chandhok, James J. Little et al.

Vision-Language Models (VLMs) have achieved impressive performance across a wide range of multimodal tasks, yet they often exhibit inconsistent behavior when faced with semantically equivalent inputs, undermining their reliability and robustness. Recent benchmarks, such as MM-R3, highlight that even state-of-the-art VLMs can produce divergent predictions across semantically equivalent inputs, despite maintaining high average accuracy. Prior work addresses this issue by modifying model architectures or conducting large-scale fine-tuning on curated datasets. In contrast, we propose a simple and effective test-time consistency framework that enhances semantic consistency without supervised re-training. Our method is entirely post-hoc, model-agnostic, and applicable to any VLM with access to its weights. Given a single test point, we enforce consistent predictions via two complementary objectives: (i) a Cross-Entropy Agreement Loss that aligns predictive distributions across semantically equivalent inputs, and (ii) a Pseudo-Label Consistency Loss that draws outputs toward a self-averaged consensus. Our method is plug-and-play and leverages information from a single test input itself to improve consistency. Experiments on the MM-R3 benchmark show that our framework yields substantial gains in consistency across state-of-the-art models, establishing a new direction for inference-time adaptation in multimodal learning.

Hari Chandana Kuchibhotla, Sumitra S Malagi, Shivam Chandhok et al.

Recent progress towards learning from limited supervision has encouraged efforts towards designing models that can recognize novel classes at test time (generalized zero-shot learning or GZSL). GZSL approaches assume knowledge of all classes, with or without labeled data, beforehand. However, practical scenarios demand models that are adaptable and can handle dynamic addition of new seen and unseen classes on the fly (that is continual generalized zero-shot learning or CGZSL). One solution is to sequentially retrain and reuse conventional GZSL methods, however, such an approach suffers from catastrophic forgetting leading to suboptimal generalization performance. A few recent efforts towards tackling CGZSL have been limited by difference in settings, practicality, data splits and protocols followed-inhibiting fair comparison and a clear direction forward. Motivated from these observations, in this work, we firstly consolidate the different CGZSL setting variants and propose a new Online-CGZSL setting which is more practical and flexible. Secondly, we introduce a unified feature-generative framework for CGZSL that leverages bi-directional incremental alignment to dynamically adapt to addition of new classes, with or without labeled data, that arrive over time in any of these CGZSL settings. Our comprehensive experiments and analysis on five benchmark datasets and comparison with baselines show that our approach consistently outperforms existing methods, especially on the more practical Online setting.

Puneet Mangla, Shivam Chandhok, Vineeth N Balasubramanian et al.

Recent progress towards designing models that can generalize to unseen domains (i.e domain generalization) or unseen classes (i.e zero-shot learning) has embarked interest towards building models that can tackle both domain-shift and semantic shift simultaneously (i.e zero-shot domain generalization). For models to generalize to unseen classes in unseen domains, it is crucial to learn feature representation that preserves class-level (domain-invariant) as well as domain-specific information. Motivated from the success of generative zero-shot approaches, we propose a feature generative framework integrated with a COntext COnditional Adaptive (COCOA) Batch-Normalization to seamlessly integrate class-level semantic and domain-specific information. The generated visual features better capture the underlying data distribution enabling us to generalize to unseen classes and domains at test-time. We thoroughly evaluate and analyse our approach on established large-scale benchmark - DomainNet and demonstrate promising performance over baselines and state-of-art methods.

Shivam Chandhok, Sanath Narayan, Hisham Cholakkal et al.

The need to address the scarcity of task-specific annotated data has resulted in concerted efforts in recent years for specific settings such as zero-shot learning (ZSL) and domain generalization (DG), to separately address the issues of semantic shift and domain shift, respectively. However, real-world applications often do not have constrained settings and necessitate handling unseen classes in unseen domains -- a setting called Zero-shot Domain Generalization, which presents the issues of domain and semantic shifts simultaneously. In this work, we propose a novel approach that learns domain-agnostic structured latent embeddings by projecting images from different domains as well as class-specific semantic text-based representations to a common latent space. In particular, our method jointly strives for the following objectives: (i) aligning the multimodal cues from visual and text-based semantic concepts; (ii) partitioning the common latent space according to the domain-agnostic class-level semantic concepts; and (iii) learning a domain invariance w.r.t the visual-semantic joint distribution for generalizing to unseen classes in unseen domains. Our experiments on the challenging DomainNet and DomainNet-LS benchmarks show the superiority of our approach over existing methods, with significant gains on difficult domains like quickdraw and sketch.

Gaurav Bhatt, Shivam Chandhok, Vineeth N Balasubramanian

A common problem with most zero and few-shot learning approaches is they suffer from bias towards seen classes resulting in sub-optimal performance. Existing efforts aim to utilize unlabeled images from unseen classes (i.e transductive zero-shot) during training to enable generalization. However, this limits their use in practical scenarios where data from target unseen classes is unavailable or infeasible to collect. In this work, we present a practical setting of inductive zero and few-shot learning, where unlabeled images from other out-of-data classes, that do not belong to seen or unseen categories, can be used to improve generalization in any-shot learning. We leverage a formulation based on product-of-experts and introduce a new AUD module that enables us to use unlabeled samples from out-of-data classes which are usually easily available and practically entail no annotation cost. In addition, we also demonstrate the applicability of our model to address a more practical and challenging, Generalized Zero-shot under a limited supervision setting, where even base seen classes do not have sufficient annotated samples.

Shivam Chandhok, Vineeth N Balasubramanian

The performance of generative zero-shot methods mainly depends on the quality of generated features and how well the model facilitates knowledge transfer between visual and semantic domains. The quality of generated features is a direct consequence of the ability of the model to capture the several modes of the underlying data distribution. To address these issues, we propose a new two-level joint maximization idea to augment the generative network with an inference network during training which helps our model capture the several modes of the data and generate features that better represent the underlying data distribution. This provides strong cross-modal interaction for effective transfer of knowledge between visual and semantic domains. Furthermore, existing methods train the zero-shot classifier either on generate synthetic image features or latent embeddings produced by leveraging representation learning. In this work, we unify these paradigms into a single model which in addition to synthesizing image features, also utilizes the representation learning capabilities of the inference network to provide discriminative features for the final zero-shot recognition task. We evaluate our approach on four benchmark datasets i.e. CUB, FLO, AWA1 and AWA2 against several state-of-the-art methods, and show its performance. We also perform ablation studies to analyze and understand our method more carefully for the Generalized Zero-shot Learning task.

Shivam Chandhok, Himani Joshi, A V Subramanyam et al.

Introduction of spectrum-sharing in 5G and subsequent generation networks demand base-station(s) with the capability to characterize the wideband spectrum spanned over licensed, shared and unlicensed non-contiguous frequency bands. Spectrum characterization involves the identification of vacant bands along with center frequency and parameters (energy, modulation, etc.) of occupied bands. Such characterization at Nyquist sampling is area and power-hungry due to the need for high-speed digitization. Though sub-Nyquist sampling (SNS) offers an excellent alternative when the spectrum is sparse, it suffers from poor performance at low signal to noise ratio (SNR) and demands careful design and integration of digital reconstruction, tunable channelizer and characterization algorithms. In this paper, we propose a novel deep-learning framework via a single unified pipeline to accomplish two tasks: 1)~Reconstruct the signal directly from sub-Nyquist samples, and 2)~Wideband spectrum characterization. The proposed approach eliminates the need for complex signal conditioning between reconstruction and characterization and does not need complex tunable channelizers. We extensively compare the performance of our framework for a wide range of modulation schemes, SNR and channel conditions. We show that the proposed framework outperforms existing SNS based approaches and characterization performance approaches to Nyquist sampling-based framework with an increase in SNR. Easy to design and integrate along with a single unified deep learning framework make the proposed architecture a good candidate for reconfigurable platforms.