Search

Graph Isomorphism (GI) is a fundamental algorithmic problem. Amongst graph classes for which the computational complexity of GI has been resolved, trees are arguably the most fundamental. Tree Isomorphism is complete for deterministic logspace, a tiny subclass of polynomial time, by Lindell's result. Over three decades ago, he devised a deterministic logspace algorithm that computes a string which is a canon for the input tree -- two trees are isomorphic precisely when their canons are identical. Inspired by Miller-Reif's reduction of Tree Isomorphism to Polynomial Identity Testing, we present a new logspace algorithm for tree canonization fundamentally different from Lindell's algorithm. Our algorithm computes a univariate polynomial as canon for an input tree, based on the classical Eisenstein's criterion for the irreducibility of univariate polynomials. This can be implemented in logspace by invoking the well known Buss et al. algorithm for arithmetic formula evaluation. This algorithm is conceptually very simple, avoiding the delicate case analysis and complex recursion that constitute the core of Lindell's algorithm. We illustrate the adaptability of our algorithm by extending it to a couple of other classes of graphs.

This study investigates thermal transport across nanocrystalline diamond/AlGaN interfaces, crucial for enhancing thermal management in AlGaN/AlGaN-based devices. Chemical vapor deposition growth of diamond directly on AlGaN resulted in a disordered interface with a high thermal boundary resistance (TBR) of 20.6 m^2-K/GW. We employed sputtered carbide interlayers (e.g., $B_4C$, $SiC$, $B_4C/SiC$) to reduce thermal boundary resistance in diamond/AlGaN interfaces. The carbide interlayers resulted in record-low thermal boundary resistance values of 3.4 and 3.7 m^2-K/GW for Al$_{0.65}$Ga$_{0.35}$N samples with $B_4C$ and $SiC$ interlayers, respectively. STEM imaging of the interface reveals interlayer thicknesses between 1.7-2.5 nm, with an amorphous structure. Additionally, Fast-Fourier Transform (FFT) characterization of sections of the STEM images displayed sharp crystalline fringes in the AlGaN layer, confirming it was properly protected from damage from hydrogen plasma during the diamond growth. In order to accurately measure the thermal boundary resistance we develop a hybrid technique, combining time-domain thermoreflectance and steady-state thermoreflectance fitting, offering superior sensitivity to buried thermal resistances. Our findings underscore the efficacy of interlayer engineering in enhancing thermal transport and demonstrate the importance of innovative measurement techniques in accurately characterizing complex thermal interfaces. This study provides a foundation for future research in improving thermal properties of semiconductor devices through interface engineering and advanced measurement methodologies.

Computing-in-memory (CIM) promises to alleviate the Von Neumann bottleneck and accelerate data-intensive applications. Depending on the underlying technology and configuration, CIM enables implementing compute primitives in place, such as multiplication, search operations, and bulk bitwise logic operations. Emerging nonvolatile memory technologies such as spintronic Racetrack memory (RTM) promise not only unprecedented density but also significant parallelism through CIM. However, most CIM designs, including those based on RTM, exhibit high fault rates. Existing error correction codes (ECC) are not homomorphic over bitwise operations such as AND and OR, and hence cannot protect against CIM faults. This paper proposes CIRM-ECC, a technique to protect spintronic RTMs against CIM faults. At the core of CIRM-ECC, we use a recently proposed RTM-based CIM approach and leverage its peripheral circuitry to our implement our novel ECC codes. We show that CIRM-ECC can be applied to single-bit Hamming codes as well as multi-bit BCH codes., Comment: 4 pages, 6 figures, to be submitted to IEEE CAL

Which dynamic queries can be maintained efficiently? For constant-size changes, it is known that constant-depth circuits or, equivalently, first-order updates suffice for maintaining many important queries, among them reachability, tree isomorphism, and the word problem for context-free languages. In other words, these queries are in the dynamic complexity class DynFO. We show that most of the existing results for constant-size changes can be recovered for batch changes of polylogarithmic size if one allows circuits of depth O(log log n) or, equivalently, first-order updates that are iterated O(log log n) times., Comment: Full version of an MFCS 2024 paper

Ferroelectricity is characterized by the presence of spontaneous and switchable macroscopic polarization. Scaling limits of ferroelectricity have been of both fundamental and technological importance, but the probes of ferroelectricity have often been indirect due to confounding factors such as leakage in the direct electrical measurements. Recent interest in low-voltage switching electronic devices squarely puts the focus on ultrathin limits of ferroelectricity in an electronic device form, specifically on the robustness of ferroelectric characteristics such as retention and endurance for practical applications. Here, we illustrate how manipulating the kinetic energy of the plasma plume during pulsed laser deposition can yield ultrathin ferroelectric capacitor heterostructures with high bulk and interface quality, significantly low leakage currents and a broad "growth window". These heterostructures venture into previously unexplored aspects of ferroelectric properties, showcasing ultralow switching voltages ($<$0.3 V), long retention times ($>$10$^{4}$s), and high endurance ($>$10$^{11}$cycles) in 20 nm films of the prototypical perovskite ferroelectric, BaTiO$_{3}$. Our work demonstrates that materials engineering can push the envelope of performance for ferroelectric materials and devices at the ultrathin limit and opens a direct, reliable and scalable pathway to practical applications of ferroelectrics in ultralow voltage switches for logic and memory technologies.

High voltage (~2 kV) AlGaN-channel HEMTs were fabricated with 64% Aluminum composition in the channel. The average on-resistance was ~75 ohm. mm (~21 miliohm. cm^2) for LGD = 20 microns. Breakdown voltage reached >3 kV (tool limit) before passivation however it reduced to ~2 kV after SiN surface passivation and field plates. The apparent high breakdown voltage prior to passivation can possibly be attributed to the field plate effect of the charged trap states of the surface. The breakdown voltage and RON demonstrated a strong linear correlation in a scattered plot with ~50 measured transistors. In pulsed IV measurements with 100 microsecond pulse width and 40 V of off-state bias (tool limit), the dynamic RON increased by ~5% compared to DC RON and current collapse was <10%., Comment: 4 pages, 6 figures

In real-time status update services for the Internet of Things (IoT), the timely dissemination of information requiring timely updates is crucial to maintaining its relevance. Failing to keep up with these updates results in outdated information. The age of information (AoI) serves as a metric to quantify the freshness of information. The Existing works to optimize AoI primarily focus on the transmission time from the information source to the monitor, neglecting the transmission time from the monitor to the destination. This oversight significantly impacts information freshness and subsequently affects decision-making accuracy. To address this gap, we designed an edge-enabled vehicular fog system to lighten the computational burden on IoT devices. We examined how information transmission and request-response times influence end-to-end AoI. As a solution, we proposed Dueling-Deep Queue Network (dueling-DQN), a deep reinforcement learning (DRL)-based algorithm and compared its performance with DQN policy and analytical results. Our simulation results demonstrate that the proposed dueling-DQN algorithm outperforms both DQN and analytical methods, highlighting its effectiveness in improving real-time system information freshness. Considering the complete end-to-end transmission process, our optimization approach can improve decision-making performance and overall system efficiency.

The emergence of 5G networks has enabled the deployment of a two-tier edge and vehicular-fog network. It comprises Multi-access Edge Computing (MEC) and Vehicular-Fogs (VFs), strategically positioned closer to Internet of Things (IoT) devices, reducing propagation latency compared to cloud-based solutions and ensuring satisfactory quality of service (QoS). However, during high-traffic events like concerts or athletic contests, MEC sites may face congestion and become overloaded. Utilizing offloading techniques, we can transfer computationally intensive tasks from resource-constrained devices to those with sufficient capacity, for accelerating tasks and extending device battery life. In this research, we consider offloading within a two-tier MEC and VF architecture, involving offloading from MEC to MEC and from MEC to VF. The primary objective is to minimize the average system cost, considering both latency and energy consumption. To achieve this goal, we formulate a multi-objective optimization problem aimed at minimizing latency and energy while considering given resource constraints. To facilitate decision-making for nearly optimal computational offloading, we design an equivalent reinforcement learning environment that accurately represents the network architecture and the formulated problem. To accomplish this, we propose a Distributed-TD3 (DTD3) approach, which builds on the TD3 algorithm. Extensive simulations, demonstrate that our strategy achieves faster convergence and higher efficiency compared to other benchmark solutions.

This paper reviews the NTIRE 2024 challenge on image super-resolution ($\times$4), highlighting the solutions proposed and the outcomes obtained. The challenge involves generating corresponding high-resolution (HR) images, magnified by a factor of four, from low-resolution (LR) inputs using prior information. The LR images originate from bicubic downsampling degradation. The aim of the challenge is to obtain designs/solutions with the most advanced SR performance, with no constraints on computational resources (e.g., model size and FLOPs) or training data. The track of this challenge assesses performance with the PSNR metric on the DIV2K testing dataset. The competition attracted 199 registrants, with 20 teams submitting valid entries. This collective endeavour not only pushes the boundaries of performance in single-image SR but also offers a comprehensive overview of current trends in this field., Comment: NTIRE 2024 webpage: https://cvlai.net/ntire/2024. Code: https://github.com/zhengchen1999/NTIRE2024_ImageSR_x4

This paper introduces a vision for Quantum Software Development lifecycle, proposing a hybrid full-stack iterative model that integrates quantum and classical computing. Addressing the current challenges in Quantum Computing (QC) such as the need for integrating diverse programming languages and managing the complexities of quantum-classical systems, this model is rooted in the principles of DevOps and continuous software engineering. It presents a comprehensive lifecycle for quantum software development, encompassing quantum-agnostic coding, testing, deployment, cloud computing services, orchestration, translation, execution, and interpretation phases. Each phase is designed to accommodate the unique demands of QC, enabling traditional software developers to engage with QC environments without needing in-depth QC expertise. The paper presents a detailed implementation roadmap, utilizing a range of existing tools and frameworks, thereby making quantum software development more accessible and efficient. The proposed model not only addresses current challenges in quantum software development but also makes a substantial contribution to the field of Quantum Software Engineering (QSE). By proposing a structured and accessible model, it sets the stage for further advancements and research in QSE, enhancing its practicality and relevance in a wide range of applications.

Mesoporous silica particles are promising candidates for drug delivery applications. In this paper, we first synthesize mesoporous silica MCM-41 and its derivative MCM-41GA with anchored glutaraldehyde bridges, and characterize them using a variety of techniques, including nitrogen adsorption/desorption, X-ray diffraction, NMR spectroscopy, scanning electron microscopy, and thermogravimetric analysis. Then, we employ random optical fields to sort mesoporous silica particles. Random optical fields by containing local intensity gradients throughout a wide range of field of view provide an elegant, easy-to-implement, and low-cost variant of multiple optical tweezers, which is known as speckle tweezers (ST). ST, similar to multiple optical tweezers, for manipulation tasks, such as trapping, sorting, and guiding of collection of micro and sub-micro objects in several disciplines including statistical physics, chemistry, microfluidics and material science. We show that ST can restrict, sieve, and sort MCM-41 and MCM-41GA particles. The different interaction of mesoporous silica variations with the applied ST may be attributed to the pre-applied modification and the differences in the porosity structure and distribution. Therefore, the results provide insight into the textural and chemical characteristics of mesoporous materials, contributing to a deeper understanding of their potential applications.

In today's data-centric world, where data fuels numerous application domains, with machine learning at the forefront, handling the enormous volume of data efficiently in terms of time and energy presents a formidable challenge. Conventional computing systems and accelerators are continually being pushed to their limits to stay competitive. In this context, computing near-memory (CNM) and computing-in-memory (CIM) have emerged as potentially game-changing paradigms. This survey introduces the basics of CNM and CIM architectures, including their underlying technologies and working principles. We focus particularly on CIM and CNM architectures that have either been prototyped or commercialized. While surveying the evolving CIM and CNM landscape in academia and industry, we discuss the potential benefits in terms of performance, energy, and cost, along with the challenges associated with these cutting-edge computing paradigms.

The accuracy of neural networks has greatly improved across various domains over the past years. Their ever-increasing complexity, however, leads to prohibitively high energy demands and latency in von Neumann systems. Several computing-in-memory (CIM) systems have recently been proposed to overcome this, but trade-offs involving accuracy, hardware reliability, and scalability for large models remain a challenge. Additionally, for some CIM designs, the activation movement still requires considerable time and energy. This paper explores the combination of algorithmic optimizations for ternary weight neural networks and associative processors (APs) implemented using racetrack memory (RTM). We propose a novel compilation flow to optimize convolutions on APs by reducing their arithmetic intensity. By leveraging the benefits of RTM-based APs, this approach substantially reduces data transfers within the memory while addressing accuracy, energy efficiency, and reliability concerns. Concretely, our solution improves the energy efficiency of ResNet-18 inference on ImageNet by 7.5x compared to crossbar in-memory accelerators while retaining software accuracy., Comment: To be presented at DATE24

Machine learning and data analytics applications increasingly suffer from the high latency and energy consumption of conventional von Neumann architectures. Recently, several in-memory and near-memory systems have been proposed to remove this von Neumann bottleneck. Platforms based on content-addressable memories (CAMs) are particularly interesting due to their efficient support for the search-based operations that form the foundation for many applications, including K-nearest neighbors (KNN), high-dimensional computing (HDC), recommender systems, and one-shot learning among others. Today, these platforms are designed by hand and can only be programmed with low-level code, accessible only to hardware experts. In this paper, we introduce C4CAM, the first compiler framework to quickly explore CAM configurations and to seamlessly generate code from high-level TorchScript code. C4CAM employs a hierarchy of abstractions that progressively lowers programs, allowing code transformations at the most suitable abstraction level. Depending on the type and technology, CAM arrays exhibit varying latencies and power profiles. Our framework allows analyzing the impact of such differences in terms of system-level performance and energy consumption, and thus supports designers in selecting appropriate designs for a given application., Comment: 10 pages, 9 figures

Let $G$ be a finite group given as input by its multiplication table. For a subset $S$ of $G$ and an element $g\in G$ the Cayley Group Membership Problem (denoted CGM) is to check if $g$ belongs to the subgroup generated by $S$. While this problem is easily seen to be in polynomial time, pinpointing its parallel complexity has been of research interest over the years. In this paper we further explore the parallel complexity of the abelian CGM problem, with focus on the dynamic setting: the generating set $S$ changes with insertions and deletions and the goal is to maintain a data structure that supports efficient membership queries to the subgroup $\angle{S}$. We obtain the following results: 1. We first consider the more general problem of Monoid Membership. When $G$ is a commutative monoid we give a deterministic dynamic algorithm constant time parallel algorithm for membership testing that supports $O(1)$ insertions and deletions in each step. 2. Building on the previous result we show that there is a dynamic randomized constant-time parallel algorithm for abelian CGM that supports polylogarithmically many insertions/deletions to $S$ in each step. 3. If the number of insertions/deletions is at most $O(\log n/\log\log n)$ then we obtain a deterministic dynamic constant-time parallel algorithm for the problem. 4. We obtain analogous results for the dynamic abelian Group Isomorphism.

Learning node-level representations of heterophilic graphs is crucial for various applications, including fraudster detection and protein function prediction. In such graphs, nodes share structural similarity identified by the equivalence of their connectivity which is implicitly encoded in the form of higher-order hierarchical information in the graphs. The contrastive methods are popular choices for learning the representation of nodes in a graph. However, existing contrastive methods struggle to capture higher-order graph structures. To address this limitation, we propose a novel multiview contrastive learning approach that integrates diffusion filters on graphs. By incorporating multiple graph views as augmentations, our method captures the structural equivalence in heterophilic graphs, enabling the discovery of hidden relationships and similarities not apparent in traditional node representations. Our approach outperforms baselines on synthetic and real structural datasets, surpassing the best baseline by $16.06\%$ on Cornell, $3.27\%$ on Texas, and $8.04\%$ on Wisconsin. Additionally, it consistently achieves superior performance on proximal tasks, demonstrating its effectiveness in uncovering structural information and improving downstream applications.

Planar Embedding is a drawing of a graph on the plane such that the edges do not intersect each other except at the vertices. We know that testing the planarity of a graph and computing its embedding (if it exists), can efficiently be computed, both sequentially [HT] and in parallel [RR94], when the entire graph is presented as input. In the dynamic setting, the input graph changes one edge at a time through insertion and deletions and planarity testing/embedding has to be updated after every change. By storing auxilliary information we can improve the complexity of dynamic planarity testing/embedding over the obvious recomputation from scratch. In the sequential dynamic setting, there has been a series of works [EGIS, IPR, HIKLR, HR1], culminating in the breakthrough result of polylog(n) sequential time (amortized) planarity testing algorithm of Holm and Rotenberg [HR2]. In this paper, we study planar embedding through the lens of DynFO, a parallel dynamic complexity class introduced by Patnaik et al. [PI] (also [DST95]). We show that it is possible to dynamically maintain whether an edge can be inserted to a planar graph without causing non-planarity in DynFO. We extend this to show how to maintain an embedding of a planar graph under both edge insertions and deletions, while rejecting edge insertions that violate planarity. Our main idea is to maintain embeddings of only the triconnected components and a special two-colouring of separating pairs that enables us to side-step cascading flips when embedding of a biconnected planar graph changes, a major issue for sequential dynamic algorithms [HR1, HR2]., Comment: To appear at MFCS 2023

The rise of data-intensive applications exposed the limitations of conventional processor-centric von-Neumann architectures that struggle to meet the off-chip memory bandwidth demand. Therefore, recent innovations in computer architecture advocate compute-in-memory (CIM) and compute-near-memory (CNM), non-von- Neumann paradigms achieving orders-of-magnitude improvements in performance and energy consumption. Despite significant technological breakthroughs in the last few years, the programmability of these systems is still a serious challenge. Their programming models are too low-level and specific to particular system implementations. Since such future architectures are predicted to be highly heterogenous, developing novel compiler abstractions and frameworks become necessary. To this end, we present CINM (Cinnamon), a first end-to-end compilation flow that leverages the hierarchal abstractions to generalize over different CIM and CNM devices and enable device-agnostic and device-aware optimizations. Cinnamon progressively lowers input programs and performs optimizations at each level in the lowering pipeline. To show its efficacy, we evaluate CINM on a set of benchmarks for the well-known UPMEM CNM system and the memristors-based CIM accelerators. We show that Cinnamon, supporting multiple hardware targets, generates high-performance code comparable to or better than state-of-the-art implementations., Comment: 16 pages, 12 figures

Optimizing combinatorial structures is core to many real-world problems, such as those encountered in life sciences. For example, one of the crucial steps involved in antibody design is to find an arrangement of amino acids in a protein sequence that improves its binding with a pathogen. Combinatorial optimization of antibodies is difficult due to extremely large search spaces and non-linear objectives. Even for modest antibody design problems, where proteins have a sequence length of eleven, we are faced with searching over 2.05 x 10^14 structures. Applying traditional Reinforcement Learning algorithms such as Q-learning to combinatorial optimization results in poor performance. We propose Structured Q-learning (SQL), an extension of Q-learning that incorporates structural priors for combinatorial optimization. Using a molecular docking simulator, we demonstrate that SQL finds high binding energy sequences and performs favourably against baselines on eight challenging antibody design tasks, including designing antibodies for SARS-COV.

In an unsupervised attack on variational autoencoders (VAEs), an adversary finds a small perturbation in an input sample that significantly changes its latent space encoding, thereby compromising the reconstruction for a fixed decoder. A known reason for such vulnerability is the distortions in the latent space resulting from a mismatch between approximated latent posterior and a prior distribution. Consequently, a slight change in an input sample can move its encoding to a low/zero density region in the latent space resulting in an unconstrained generation. This paper demonstrates that an optimal way for an adversary to attack VAEs is to exploit a directional bias of a stochastic pullback metric tensor induced by the encoder and decoder networks. The pullback metric tensor of an encoder measures the change in infinitesimal latent volume from an input to a latent space. Thus, it can be viewed as a lens to analyse the effect of input perturbations leading to latent space distortions. We propose robustness evaluation scores using the eigenspectrum of a pullback metric tensor. Moreover, we empirically show that the scores correlate with the robustness parameter $\beta$ of the $\beta-$VAE. Since increasing $\beta$ also degrades reconstruction quality, we demonstrate a simple alternative using \textit{mixup} training to fill the empty regions in the latent space, thus improving robustness with improved reconstruction., Comment: International Conference on Artificial Intelligence and Statistics (AISTATS) 2023

Dynamic Complexity was introduced by Immerman and Patnaik PI97 in the nineties and has seen a resurgence of interest with the positive resolution of their conjecture on directed reachability in DynFO DKMSZ18. Since then many natural problems related to reachability and matching have been placed in DynFO and related classes DMVZ18,DKMTVZ20,DTV21. In this work, we place some dynamic problems from group theory in DynFO. In particular, suppose we are given an arbitrary multiplication table over n elements representing an unstructured binary operation (representing a structure called a magma). Suppose the table evolves through a change in one of its n^2 entries in one step. For a set S of magma elements which also changes one element at a time, we can maintain enough auxiliary information so that when the magma is a group, we are able to answer the Cayley Group Membership (CGM) problem for S and a target t (i.e. "Is t a product of elements from S? ") using an FO query at every step. This places the dynamic CGM problem (for groups) when the ambient magma is specified via a table in DynFO. In contrast, for the table setting, statically CGM was known to be in the class Logspace BarringtonM06. Building on the dynamic CGM result, we can maintain the isomorphism of of two magmas, whenever both are Abelian groups, in DynFO. Our techniques include a way to maintain the powers of the elements of a magma in DynFO using left associative parenthesisation, the notion of cube independence to cube generate a subgroup generated by a set, a way to maintain maximal cube independent sequences in a magma along with some group theoretic machinery available from McKenzieCook. The notion of cube independent sequences is new as far as we know and may be of independent interest. These techniques are very different from the ones employed in Dynamic Complexity so far., Comment: The paper has been withdrawn since some crucial lemmas are wrong (Theorem 1, Lemma 10) related to cube bases. Please refer to arXiv:2308.10073 for the follow up with updated results

Sequence diversity is one of the major challenges in the design of diagnostic, prophylactic and therapeutic interventions against viruses. DiMA is a novel tool that is big data-ready and designed to facilitate the dissection of sequence diversity dynamics for viruses. DiMA stands out from other diversity analysis tools by offering various unique features. DiMA provides a quantitative overview of sequence (nucleotide/protein) diversity by use of Shannon's entropy corrected for size bias, applied via a user-defined k-mer sliding window to an input alignment file, and each k-mer position is dissected to various diversity motifs. The motifs are defined based on the probability of distinct sequences at a given k-mer position, whereby an index is the predominant sequence, while all the others are (total) variants to the index. The total variants are sub-classified into the major (most common) variant, minor variants (occurring more than once and of frequency lower than the major), and the unique (singleton) variants. DiMA allows user-defined, sequence metadata enrichment for analyses of the motifs. The application of DiMA was demonstrated for the alignment data of the relatively conserved Spike protein (2,106,985 sequences) of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and the relatively highly diverse Pol protein (3,874) of human immunodeficiency virus-1 (HIV-1). The tool is publicly available as a web server (https://dima.bezmialem.edu.tr), as a Python library (via PyPi) and as a command line client (Via GitHub)., Comment: 18 pages, 2 figures, 50 references

Recent DNA pre-alignment filter designs employ DRAM for storing the reference genome and its associated meta-data. However, DRAM incurs increasingly high energy consumption background and refresh energy as devices scale. To overcome this problem, this paper explores a design with racetrack memory (RTM)--an emerging non-volatile memory that promises higher storage density, faster access latency, and lower energy consumption. Multi-bit storage cells in RTM are inherently sequential and thus require data placement strategies to mitigate the performance and energy impacts of shifting during data accesses. We propose a near-memory pre-alignment filter with a novel data mapping and several shift reduction strategies designed explicitly for RTM. On a set of four input genomes from the 1000 Genome Project, our approach improves performance and energy efficiency by 68% and 52%, respectively, compared to the state of the art proposed DRAM-based architecture.

Phase transition between ferroelectricity and quantum paraelectricity via non-thermal tuning parameters can lead to quantum critical behavior and associated emergent phenomena. Ferroelectric quantum critical systems are, however, rare despite the abundance of ferroelectric materials. Here, we show theoretically that in ferroelectric-paraelectric heterostructures, it is plausible to induce quantum paraelectricity where the quantum temperature (i.e., the temperature below with the onset of ferroelectricity is suppressed by quantum fluctuations) can be tuned by the thickness ratio. This, in turn, can effect a quantum phase transition between effective ferroelectric and quantum paraelectric states, using the thickness ratio as the tuning parameter. The associated quantum critical region offers unexpected prospects in the field of ferroelectric quantum criticality., Comment: Manuscript: 16 pages including 5 figures Supplementary Material: 7 pages including 3 figures

Antibodies are canonically Y-shaped multimeric proteins capable of highly specific molecular recognition. The CDRH3 region located at the tip of variable chains of an antibody dominates antigen-binding specificity. Therefore, it is a priority to design optimal antigen-specific CDRH3 regions to develop therapeutic antibodies. However, the combinatorial nature of CDRH3 sequence space makes it impossible to search for an optimal binding sequence exhaustively and efficiently using computational approaches. Here, we present \texttt{AntBO}: a combinatorial Bayesian optimisation framework enabling efficient \textit{in silico} design of the CDRH3 region. Ideally, antibodies are expected to have high target specificity and developability. We introduce a CDRH3 trust region that restricts the search to sequences with favourable developability scores to achieve this goal. For benchmarking, \texttt{AntBO} uses the \texttt{Absolut!} software suite as a black-box oracle to score the target specificity and affinity of designed antibodies \textit{in silico} in an unconstrained fashion~\citep{robert2021one}. The experiments performed for $159$ discretised antigens used in \texttt{Absolut!} demonstrate the benefit of \texttt{AntBO} in designing CDRH3 regions with diverse biophysical properties. In under $200$ calls to black-box oracle, \texttt{AntBO} can suggest antibody sequences that outperform the best binding sequence drawn from 6.9 million experimentally obtained CDRH3s and a commonly used genetic algorithm baseline. Additionally, \texttt{AntBO} finds very-high affinity CDRH3 sequences in only 38 protein designs whilst requiring no domain knowledge. We conclude \texttt{AntBO} brings automated antibody design methods closer to what is practically viable for in vitro experimentation.