Sourav Das

Luigi Filippo Chiara, Pasquale Coscia, Sourav Das et al.

Human trajectory forecasting is a key component of autonomous vehicles, social-aware robots and advanced video-surveillance applications. This challenging task typically requires knowledge about past motion, the environment and likely destination areas. In this context, multi-modality is a fundamental aspect and its effective modeling can be beneficial to any architecture. Inferring accurate trajectories is nevertheless challenging, due to the inherently uncertain nature of the future. To overcome these difficulties, recent models use different inputs and propose to model human intentions using complex fusion mechanisms. In this respect, we propose a lightweight attention-based recurrent backbone that acts solely on past observed positions. Although this backbone already provides promising results, we demonstrate that its prediction accuracy can be improved considerably when combined with a scene-aware goal-estimation module. To this end, we employ a common goal module, based on a U-Net architecture, which additionally extracts semantic information to predict scene-compliant destinations. We conduct extensive experiments on publicly-available datasets (i.e. SDD, inD, ETH/UCY) and show that our approach performs on par with state-of-the-art techniques while reducing model complexity.

Sourav Das

Small Language Models (SLMs) offer a balance between capability and computational feasibility. Neural scaling laws inform their optimal training, suggesting that they possess rich internal representations that scale with their size. However, deploying even these SLMs can be challenging under strict resource constraints. Language model probing provides methods for analyzing the linguistic knowledge encoded in a model's internals. We propose ProbScale, a framework that unifies insights from scaling laws and probing to identify parameter-efficient subnetworks within pre-trained SLMs. ProbScale utilizes the high-quality representations of well-scaled SLMs and uses task-specific probes to mathematically quantify the relevance of each layer for target downstream capabilities. This allows selecting subnetworks that optimally trade off performance against parameter size. We formulate the subnetwork selection as finding a layer subset maximizing aggregated, task-weighted probe performance under a parameter budget. Experiments on representative SLMs such as RoBERTa-Large and T5-Base demonstrate that ProbScale identifies subnetworks achieving significant parameter reduction, from 5 to 10 times, while maintaining high performance (95% to 98% of the original SLMs) on targeted tasks, outperforming heuristic baselines.

Saptarshi Das, Sourav Das, Koushik Maharatna

In this paper, an efficient control strategy for physiological interaction based anaesthetic drug infusion model is explored using the fractional order (FO) proportional integral derivative (PID) controllers. The dynamic model is composed of several human organs by considering the brain response to the anaesthetic drug as output and the drug infusion rate as the control input. Particle Swarm Optimisation (PSO) is employed to obtain the optimal set of parameters for PID/FOPID controller structures. With the proposed FOPID control scheme much less amount of drug-infusion system can be designed to attain a specific anaesthetic target and also shows high robustness for +/-50% parametric uncertainty in the patient's brain model.

Avijit Routh, Sourav Das, Saptarshi Das et al.

Stabilizing state feedback controller has been designed in this paper for a switched DC motor plant, controlled over communication network. The switched system formulation for the networked control system (NCS) with additional switching in a plant parameter along with the switching due to random packet losses, have been formulated as few set of non-strict Linear Matrix Inequalities (LMIs). In order to solve non-strict LMIs using standard LMI solver and to design the stabilizing state feedback controller, the Cone Complementary Linearization (CCL) technique has been adopted. Simulation studies have been carried out for a DC motor plant, operating at two different sampling times with random switching in the moment of inertia, representing sudden jerks.

Mohammad Mussadiq Jalalzai, Kushal Babel, Jovan Komatovic et al.

This paper introduces MonadBFT, a novel Byzantine Fault Tolerant (BFT) consensus protocol that advances both performance and robustness. MonadBFT is implemented as the consensus protocol in the Monad blockchain. As a HotStuff-family protocol, MonadBFT has linear message complexity in the common case and is optimistically responsive, operating as quickly as the network allows. A central feature of MonadBFT is its tail-forking resistance. In pipelined BFT protocols, when a leader goes offline, the previous proposal is abandoned. Malicious leaders can exploit this tail-forking behavior as a form of Maximal Extractable Value (MEV) attack by deliberately discarding their predecessor's block, depriving that proposer of rewards and enabling transaction reordering, censorship or theft. MonadBFT prevents such tail-forking attacks, preserving both fairness and integrity in transaction execution. Another related feature of MonadBFT is its notion of speculative finality, which enables parties to execute ordered transactions after a single round (i.e., a single view), with reverts occurring only in the rare case of provable leader equivocation. This mechanism reduces user-perceived latency. Additionally, we introduce the leader fault isolation property, which ensures that the protocol can quickly recover from a failure. To our knowledge, no prior pipelined, leader-based BFT consensus protocol combines all of these properties in a single design.

Sourav Das, Anup Kumar Kolya

With the ever-increasing number of digital music and vast music track features through popular online music streaming software and apps, feature recognition using the neural network is being used for experimentation to produce a wide range of results across a variety of experiments recently. Through this work, the authors extract information on such features from a popular open-source music corpus and explored new recognition techniques, by applying unsupervised Hebbian learning techniques on their single-layer neural network using the same dataset. The authors show the detailed empirical findings to simulate how such an algorithm can help a single layer feedforward network in training for music feature learning as patterns. The unsupervised training algorithm enhances their proposed neural network to achieve an accuracy of 90.36% for successful music feature detection. For comparative analysis against similar tasks, authors put their results with the likes of several previous benchmark works. They further discuss the limitations and thorough error analysis of their work. The authors hope to discover and gather new information about this particular classification technique and its performance, and further understand future potential directions and prospects that could improve the art of computational music feature recognition.

Sourav Das, Guglielmo Camporese, Shaokang Cheng et al.

Long-term trajectory forecasting is an important and challenging problem in the fields of computer vision, machine learning, and robotics. One fundamental difficulty stands in the evolution of the trajectory that becomes more and more uncertain and unpredictable as the time horizon grows, subsequently increasing the complexity of the problem. To overcome this issue, in this paper, we propose Di-Long, a new method that employs the distillation of a short-term trajectory model forecaster that guides a student network for long-term trajectory prediction during the training process. Given a total sequence length that comprehends the allowed observation for the student network and the complementary target sequence, we let the student and the teacher solve two different related tasks defined over the same full trajectory: the student observes a short sequence and predicts a long trajectory, whereas the teacher observes a longer sequence and predicts the remaining short target trajectory. The teacher's task is less uncertain, and we use its accurate predictions to guide the student through our knowledge distillation framework, reducing long-term future uncertainty. Our experiments show that our proposed Di-Long method is effective for long-term forecasting and achieves state-of-the-art performance on the Intersection Drone Dataset (inD) and the Stanford Drone Dataset (SDD).

Sourav Das, Solomon Tesfamariam

This paper presents a comprehensive review of the design of experiments used in the surrogate models. In particular, this study demonstrates the necessity of the design of experiment schemes for the Physics-Informed Neural Network (PINN), which belongs to the supervised learning class. Many complex partial differential equations (PDEs) do not have any analytical solution; only numerical methods are used to solve the equations, which is computationally expensive. In recent decades, PINN has gained popularity as a replacement for numerical methods to reduce the computational budget. PINN uses physical information in the form of differential equations to enhance the performance of the neural networks. Though it works efficiently, the choice of the design of experiment scheme is important as the accuracy of the predicted responses using PINN depends on the training data. In this study, five different PDEs are used for numerical purposes, i.e., viscous Burger's equation, Shrödinger equation, heat equation, Allen-Cahn equation, and Korteweg-de Vries equation. A comparative study is performed to establish the necessity of the selection of a DoE scheme. It is seen that the Hammersley sampling-based PINN performs better than other DoE sample strategies.

Conor O'Brien, Arvind Thiagarajan, Sourav Das et al.

Online advertising has typically been more personalized than offline advertising, through the use of machine learning models and real-time auctions for ad targeting. One specific task, predicting the likelihood of conversion (i.e.\ the probability a user will purchase the advertised product), is crucial to the advertising ecosystem for both targeting and pricing ads. Currently, these models are often trained by observing individual user behavior, but, increasingly, regulatory and technical constraints are requiring privacy-preserving approaches. For example, major platforms are moving to restrict tracking individual user events across multiple applications, and governments around the world have shown steadily more interest in regulating the use of personal data. Instead of receiving data about individual user behavior, advertisers may receive privacy-preserving feedback, such as the number of installs of an advertised app that resulted from a group of users. In this paper we outline the recent privacy-related changes in the online advertising ecosystem from a machine learning perspective. We provide an overview of the challenges and constraints when learning conversion models in this setting. We introduce a novel approach for training these models that makes use of post-ranking signals. We show using offline experiments on real world data that it outperforms a model relying on opt-in data alone, and significantly reduces model degradation when no individual labels are available. Finally, we discuss future directions for research in this evolving area.

Sourav Das, Prasanta Mandal, Sanjay Chatterji

Social media has seen a worrying rise in hate speech in recent times. Branching to several distinct categories of cyberbullying, gender discrimination, or racism, the combined label for such derogatory content can be classified as toxic content in general. This paper presents experimentation with a Keras wrapped lightweight BERT model to successfully identify hate speech and predict probabilistic impact score for the same to extract the hateful words within sentences. The dataset used for this task is the Hate Speech and Offensive Content Detection (HASOC 2021) data from FIRE 2021 in English. Our system obtained a validation accuracy of 82.60%, with a maximum F1-Score of 82.68%. Subsequently, our predictive cases performed significantly well in generating impact scores for successful identification of the hate tweets as well as the hateful words from tweet pools.

Shubham Anand Jain, Rohan Shah, Sanit Gupta et al.

We consider the problem of correctly identifying the \textit{mode} of a discrete distribution $\mathcal{P}$ with sufficiently high probability by observing a sequence of i.i.d. samples drawn from $\mathcal{P}$. This problem reduces to the estimation of a single parameter when $\mathcal{P}$ has a support set of size $K = 2$. After noting that this special case is tackled very well by prior-posterior-ratio (PPR) martingale confidence sequences \citep{waudby-ramdas-ppr}, we propose a generalisation to mode estimation, in which $\mathcal{P}$ may take $K \geq 2$ values. To begin, we show that the "one-versus-one" principle to generalise from $K = 2$ to $K \geq 2$ classes is more efficient than the "one-versus-rest" alternative. We then prove that our resulting stopping rule, denoted PPR-1v1, is asymptotically optimal (as the mistake probability is taken to $0$). PPR-1v1 is parameter-free and computationally light, and incurs significantly fewer samples than competitors even in the non-asymptotic regime. We demonstrate its gains in two practical applications of sampling: election forecasting and verification of smart contracts in blockchains.

Arunava Kumar Chakraborty, Sourav Das, Anup Kumar Kolya

As the Covid-19 outbreaks rapidly all over the world day by day and also affects the lives of million, a number of countries declared complete lock-down to check its intensity. During this lockdown period, social media plat-forms have played an important role to spread information about this pandemic across the world, as people used to express their feelings through the social networks. Considering this catastrophic situation, we developed an experimental approach to analyze the reactions of people on Twitter taking into ac-count the popular words either directly or indirectly based on this pandemic. This paper represents the sentiment analysis on collected large number of tweets on Coronavirus or Covid-19. At first, we analyze the trend of public sentiment on the topics related to Covid-19 epidemic using an evolutionary classification followed by the n-gram analysis. Then we calculated the sentiment ratings on collected tweet based on their class. Finally, we trained the long-short term network using two types of rated tweets to predict sentiment on Covid-19 data and obtained an overall accuracy of 84.46%.

Sourav Das, Anup Kumar Kolya

Sarcasm is a sophisticated way of wrapping any immanent truth, mes-sage, or even mockery within a hilarious manner. The advent of communications using social networks has mass-produced new avenues of socialization. It can be further said that humor, irony, sarcasm, and wit are the four chariots of being socially funny in the modern days. In this paper, we manually extract the sarcastic word distribution features of a benchmark pop culture sarcasm corpus, containing sarcastic dialogues and monologues. We generate input sequences formed of the weighted vectors from such words. We further propose an amalgamation of four parallel deep long-short term networks (pLSTM), each with distinctive activation classifier. These modules are primarily aimed at successfully detecting sarcasm from the text corpus. Our proposed model for detecting sarcasm peaks a training accuracy of 98.95% when trained with the discussed dataset. Consecutively, it obtains the highest of 98.31% overall validation accuracy on two handpicked Project Gutenberg English humor literature among all the test cases. Our approach transcends previous state-of-the-art works on several sarcasm corpora and results in a new gold standard performance for sarcasm detection.

Sourav Das, Vinith Krishnan, Ling Ren

Sharding is a promising blockchain scaling solution. But it currently suffers from high latency and low throughput when it comes to cross-shard transactions, i.e., transactions that require coordination from multiple shards. The root cause of these limitations arise from the use of the classic two-phase commit protocol, which involves locking assets for extended periods of time. This paper presents Rivet, a new paradigm for blockchain sharding that achieves lower latency and higher throughput for cross-shard transactions. Rivet has a single reference shard running consensus, and multiple worker shards maintaining disjoint states and processing a subset of transactions in the system. Rivet obviates the need for consensus within each worker shard, and as a result, tolerates more failures within a shard and lowers communication overhead. We prove the correctness and security of Rivet. We also propose a more realistic framework for evaluating sharded blockchains by creating a benchmark based on real Ethereum transactions. An evaluation of our prototype implementation of Rivet and the baseline two-phase commit, atop 50+ AWS EC2 instances, using our evaluation framework demonstrates the latency and throughput improvements for cross-shard transactions.

Pranay Gupta, Anirudh Thatipelli, Aditya Aggarwal et al.

In this paper, we study current and upcoming frontiers across the landscape of skeleton-based human action recognition. To study skeleton-action recognition in the wild, we introduce Skeletics-152, a curated and 3-D pose-annotated subset of RGB videos sourced from Kinetics-700, a large-scale action dataset. We extend our study to include out-of-context actions by introducing Skeleton-Mimetics, a dataset derived from the recently introduced Mimetics dataset. We also introduce Metaphorics, a dataset with caption-style annotated YouTube videos of the popular social game Dumb Charades and interpretative dance performances. We benchmark state-of-the-art models on the NTU-120 dataset and provide multi-layered assessment of the results. The results from benchmarking the top performers of NTU-120 on the newly introduced datasets reveal the challenges and domain gap induced by actions in the wild. Overall, our work characterizes the strengths and limitations of existing approaches and datasets. Via the introduced datasets, our work enables new frontiers for human action recognition.

Sourav Das, Nitin Awathare, Ling Ren et al.

Proof-of-Work~(PoW) based blockchains typically allocate only a tiny fraction (e.g., less than 1% for Ethereum) of the average interarrival time~($\mathbb{I}$) between blocks for validating transactions. A trivial increase in validation time~($τ$) introduces the popularly known Verifier's Dilemma, and as we demonstrate, causes more forking and increases unfairness. Large $τ$ also reduces the tolerance for safety against a Byzantine adversary. Solutions that offload validation to a set of non-chain nodes (a.k.a. off-chain approaches) suffer from trust issues that are non-trivial to resolve. In this paper, we present Tuxedo, the first on-chain protocol to theoretically scale $τ/\mathbb{I} \approx 1$ in PoW blockchains. The key innovation in Tuxedo is to separate the consensus on the ordering of transactions from their execution. We achieve this by allowing miners to delay validation of transactions in a block by up to $ζ$ blocks, where $ζ$ is a system parameter. We perform security analysis of Tuxedo considering all possible adversarial strategies in a synchronous network with end-to-end delay $Δ$ and demonstrate that Tuxedo achieves security equivalent to known results for longest chain PoW Nakamoto consensus. Additionally, we also suggest a principled approach for practical choices of parameter $ζ$ as per the application requirement. Our prototype implementation of Tuxedo atop Ethereum demonstrates that it can scale $τ$ without suffering the harmful effects of naive scaling in existing blockchains.

Sourav Das, Samuel Wedaj, Kolin Paul et al.

The proliferation of application specific cyber-physical systems coupled with the emergence of a variety of attacks on such systems (malware such as Mirai and Hajime) underlines the need to secure such networks. Most existing security efforts have focused on only detection of the presence of malware. However given the ability of most attacks to spread through the network once they infect a few devices, it is important to contain the spread of a virus and at the same time systematically cleanse the impacted nodes using the communication capabilities of the network. Toward this end, we present Airmed - a method and system to not just detect corruption of the application software on a IoT node, but to self correct itself using its neighbors. Airmed's decentralized mechanisms prevent the spread of self-propagating malware and can also be used as a technique for updating application code on such IoT devices. Among the novelties of Airmed are a novel bloom-filter technique along with hardware support to identify position of the malware program from the benign application code, an adaptive self-check for computational efficiency, and a uniform random-backoff and stream signatures for secure and bandwidth efficient code exchange to correct corrupted devices. We assess the performance of Airmed, using the embedded systems security architecture of TrustLite in the OMNeT++ simulator. The results show that Airmed scales up to thousands of devices, ensures guaranteed update of the entire network, and can recover 95% of the nodes in 10 minutes in both internal and external propagation models. Moreover, we evaluate memory and communication costs and show that Airmed is efficient and incurs very low overhead.

Sourav Das, Vinay Joseph Ribeiro, Abhijeet Anand

One major shortcoming of permissionless blockchains such as Bitcoin and Ethereum is that they are unsuitable for running Computationally Intensive smart Contracts (CICs). This prevents such blockchains from running Machine Learning algorithms, Zero-Knowledge proofs, etc. which may need non-trivial computation. In this paper, we present YODA, which is to the best of our knowledge the first solution for efficient computation of CICs in permissionless blockchains with guarantees for a threat model with both Byzantine and selfish nodes. YODA selects one or more execution sets (ES) via Sortition to execute a particular CIC off-chain. One key innovation is the MultI-Round Adaptive Consensus using Likelihood Estimation (MIRACLE) algorithm based on sequential hypothesis testing. M I RACLE allows the execution sets to be small thus making YODA efficient while ensuring correct CIC execution with high probability. It adapts the number of ES sets automatically depending on the concentration of Byzantine nodes in the system and is optimal in terms of the expected number of ES sets used in certain scenarios. Through a suite of economic incentives and technical mechanisms such as the novel Randomness Inserted Contract Execution (RICE) algorithm, we force selfish nodes to behave honestly. We also prove that the honest behavior of selfish nodes is an approximate Nash Equilibrium. We present the system design and details of YODA and prove the security properties of MIRACLE and RICE. Our prototype implementation built on top of Ethereum demonstrates the ability of YODA to run CICs with orders of magnitude higher gas per unit time as well as total gas requirements than Ethereum currently supports. It also demonstrates the low overheads of RICE.