Singularity

The rat kidney was peculiarly beautiful — an edgeless viscera about the size of a quarter, gemstone-like and gleaming as if encased in pure glass.

It owed its veneer to a frosty descent in liquid nitrogen vapor to minus 150-degrees Celsius, a process known as vitrification, that shocked the kidney into an icy state of suspended animation. Then researchers at the University of Minnesota restarted the kidney’s biological clock, rewarming it before transplanting it back into a live rat — who survived the ordeal.

In all, five rats received a vitrified-then-thawed kidney in a study whose results were published this month in Nature Communications. It’s the first time scientists have shown it’s possible to successfully and repeatedly transplant a life-sustaining mammalian organ after it has been rewarmed from this icy metabolic arrest. Outside experts unequivocally called the results a seminal milestone for the field of organ preservation.

Journal Article:

Vitrification and nanowarming enable long-term organ cryopreservation and life-sustaining kidney transplantation in a rat model

To accelerate development of useful new materials, researchers are building a new kind of automated lab that uses robots guided by artificial intelligence.

“Our vision is using AI to discover the materials of the future,” said Yan Zeng, a staff scientist leading the A-Lab at the Department of Energy’s Lawrence Berkeley National Laboratory (Berkeley Lab). The “A” in A-Lab is deliberately ambiguous, standing for artificial intelligence (AI), automated, accelerated, and abstracted, among others.

Scientists have computationally predicted hundreds of thousands of novel materials that could be promising for new technologies – but testing to see whether any of those materials can be made in reality is a slow process. Enter A-Lab, which can process 50 to 100 times as many samples as a human every day and use AI to quickly pursue promising finds.

A-Lab could help identify and fast-track materials for several research areas, such as solar cells, fuel cells, thermoelectrics (materials that generate energy from temperature differences), and other clean energy technologies. To start, researchers will focus on finding new materials for batteries and energy storage, addressing critical needs for an affordable, equitable, and sustainable energy supply.

For most of the history of life on Earth, genetic information has been carried in a code that specifies just 20 amino acids. Amino acids are the building blocks of proteins, which do most of the heavy lifting in the cell; their side-chains govern protein folding, interactions and chemical activities. By limiting the available side chains, nature effectively restricts the kinds of reaction that proteins can perform.

As a doctoral student in the 1980s, Peter Schultz found himself wondering why nature had restricted itself in this way — and set about trying to circumvent this limitation. Several years later, as a professor at the University of California, Berkeley, Schultz and his team managed to do so by tinkering with the machinery of protein synthesis. Although confined to a test tube, the work marked a key early success in efforts to hack the genetic code.

Since then, many researchers have followed in Schultz’s footsteps, tweaking the cellular apparatus for building proteins both to alter existing macromolecules and to create polymers from entirely new building blocks. The resulting molecules can be used in research and for the development of therapeutics and materials. But it’s been a hard slog, because protein synthesis is a crucial cellular function that cannot easily be changed.

Significance

We demonstrate the highest-resolution MR images ever obtained of the mouse brain. The diffusion tensor images (DTI) @ 15 μm spatial resolution are 1,000 times the resolution of most preclinical rodent DTI/MRI. Superresolution track density images are 27,000 times that of typical preclinical DTI/MRI. High angular resolution yielded the most detailed MR connectivity maps ever generated. High-performance computing pipelines merged the DTI with light sheet microscopy of the same specimen, providing a comprehensive picture of cells and circuits. The methods have been used to demonstrate how strain differences result in differential changes in connectivity with age. We believe the methods will have broad applicability in the study of neurodegenerative diseases.

Abstract

We have developed workflows to align 3D magnetic resonance histology (MRH) of the mouse brain with light sheet microscopy (LSM) and 3D delineations of the same specimen. We start with MRH of the brain in the skull with gradient echo and diffusion tensor imaging (DTI) at 15 μm isotropic resolution which is ~ 1,000 times higher than that of most preclinical MRI. Connectomes are generated with superresolution tract density images of ~5 μm. Brains are cleared, stained for selected proteins, and imaged by LSM at 1.8 μm/pixel. LSM data are registered into the reference MRH space with labels derived from the ABA common coordinate framework. The result is a high-dimensional integrated volume with registration (HiDiver) with alignment precision better than 50 µm. Throughput is sufficiently high that HiDiver is being used in quantitative studies of the impact of gene variants and aging on mouse brain cytoarchitecture and connectomics.

In 2023, the market for autonomous tractors is expected to be worth US$1.5 billion. The total market value is predicted to increase at a phenomenal CAGR (Compound Annual Growth Rate) of 24% from 2023 to 2033, reaching US$ 13 billion.

A team of researchers with the New York State University (NYU) has done the seemingly impossible: they've successfully designed a semiconductor chip with no hardware definition language. Using only plain English - and the definitions and examples within it that can define and describe a semiconductor processor - the team showcased what human ingenuity, curiosity, and baseline knowledge can do when aided by the AI prowess of ChatGPT.

While surprising, it goes further: the chip wasn't only designed. It was manufactured; it was benchmarked, and it worked. The two hardware engineers' usage of plain English showcases just how valuable and powerful ChatGPT can be (as if we still had doubts, following the number of awe-inspiring things it's done already).

Journal Link

Chip-Chat: Challenges and Opportunities in Conversational Hardware Design

Natural DNA is often double-stranded: one strand to encode the genes and one backup strand, intertwined in a double helix. The double helix is stabilized by Watson-Crick interactions, which allow the two strands to recognize and pair with one another. Yet there exists another, lesser-known class of interactions between DNA. These so-called normal or reverse Hoogsteen interactions allow a third strand to join in, forming a beautiful triple helix.

In a recent paper, published in Advanced Materials, researchers from the Gothelf lab debut a general method to organize double-stranded DNA, based on Hoogsteen interactions. The study unambiguously demonstrates that triplex-forming strands are capable of sharply bending or “folding” double-stranded DNA to create compacted structures. The appearance of these structures range from hollow two-dimensional shapes to dense 3D constructs and everything in-between, including a structure resembling a potted flower. Gothelf and co-workers have named their method triplex origami.

Journal Article:

Folding Double-Stranded DNA into Designed Shapes with Triplex-Forming Oligonucleotides

Abstract

Graphene has recently gained significant interest owing to its advantageous physicochemical and biological properties. However, its preparation strategies, main properties, chemical derivatives, and advanced applications in the multidimensional fields of lubrication, electricity, and tissue engineering are rarely reported. Hence, this review presents comprehensive discussions on current states of graphene as effective reinforcements to apply into these fields. First, graphene preparation methods are analyzed, and its main properties and chemical derivatives are discussed. Then, the friction-reduction and antiwear mechanisms of graphene are summarized. Next, the advanced applications of graphene in electricity and tissue engineering are described. Finally, the review is concluded by presenting outlooks on key challenges and future opportunities for extending preparation methods and multidimensional applications of the graphene-based materials.

Abstract

Advances in nanoscience have enabled the synthesis of nanomaterials, such as graphene, from low-value or waste materials through flash Joule heating. Though this capability is promising, the complex and entangled variables that govern nanocrystal formation in the Joule heating process remain poorly understood. In this work, machine learning (ML) models are constructed to explore the factors that drive the transformation of amorphous carbon into graphene nanocrystals during flash Joule heating. An XGBoost regression model of crystallinity achieves an r2 score of 0.8051 ± 0.054. Feature importance assays and decision trees extracted from these models reveal key considerations in the selection of starting materials and the role of stochastic current fluctuations in flash Joule heating synthesis. Furthermore, partial dependence analyses demonstrate the importance of charge and current density as predictors of crystallinity, implying a progression from reaction-limited to diffusion-limited kinetics as flash Joule heating parameters change. Finally, a practical application of the ML models is shown by using Bayesian meta-learning algorithms to automatically improve bulk crystallinity over many Joule heating reactions. These results illustrate the power of ML as a tool to analyze complex nanomanufacturing processes and enable the synthesis of 2D crystals with desirable properties by flash Joule heating.

Abstract

We introduce Voyager, the first LLM-powered embodied lifelong learning agent in Minecraft that continuously explores the world, acquires diverse skills, and makes novel discoveries without human intervention. Voyager consists of three key components: 1) an automatic curriculum that maximizes exploration, 2) an ever-growing skill library of executable code for storing and retrieving complex behaviors, and 3) a new iterative prompting mechanism that incorporates environment feedback, execution errors, and self-verification for program improvement. Voyager interacts with GPT-4 via blackbox queries, which bypasses the need for model parameter fine-tuning. The skills developed by Voyager are temporally extended, interpretable, and compositional, which compounds the agent's abilities rapidly and alleviates catastrophic forgetting. Empirically, Voyager shows strong in-context lifelong learning capability and exhibits exceptional proficiency in playing Minecraft. It obtains 3.3x more unique items, travels 2.3x longer distances, and unlocks key tech tree milestones up to 15.3x faster than prior SOTA. Voyager is able to utilize the learned skill library in a new Minecraft world to solve novel tasks from scratch, while other techniques struggle to generalize.

You can take a look at the technology demoed in this video at Kahnmigo!

Abstract:

CMOS technology and its continuous scaling have made electronics and computers accessible and affordable for almost everyone on the globe; in addition, they have enabled the solutions of a wide range of societal problems and applications. Today, however, both the technology and the computer architectures are facing severe challenges/walls making them incapable of providing the demanded computing power with tight constraints. This motivates the need for the exploration of novel architectures based on new device technologies; not only to sustain the financial benefit of technology scaling, but also to develop solutions for extremely demanding emerging applications. This paper presents two computation-in-memory based accelerators making use of emerging memristive devices; they are Memristive Vector Processor and RRAM Automata Processor. The preliminary results of these two accelerators show significant improvement in terms of latency, energy and area as compared to today's architectures and design.

In this paper authors from UCSB and Microsoft Research propose the LONGMEM framework, which enables language models to cache long-form prior context or knowledge into the non-differentiable memory bank and take advantage of them via a decoupled memory module to address the memory staleness problem. They create a revolutionary residual side network (SideNet) to achieve decoupled memory. A frozen backbone LLM is used to extract the paired attention keys and values from the previous context into the memory bank. The resulting attention query of the current input is utilized in the SideNet’s memory-augmented layer to access cached (keys and values) for earlier contexts. The associated memory augmentations are then fused into learning hidden states via a joint attention process.

Paper:

Augmenting Language Models with Long-Term Memory

Progress in drug testing and regenerative medicine could greatly benefit from laboratory-engineered human tissues built of a variety of cell types with precise 3D architecture. But production of greater than millimeter sized human tissues has been limited by a lack of methods for building tissues with embedded life-sustaining vascular networks.

Context is an important part of understanding the meaning of natural language, but most neuroimaging studies of meaning use isolated words and isolated sentences with little context. In this study, we examined whether the results of neuroimaging language studies that use out-of-context stimuli generalize to natural language. We find that increasing context improves the quality of neuroimaging data and changes where and how semantic information is represented in the brain. These results suggest that findings from studies using out-of-context stimuli may not generalize to natural language used in daily life

There is a major health crisis in terms of the shortage of organs. Since 2013, the total number of patients requiring a transplant has doubled while the number of available donor organs has remained relatively the same. According to the Health Resources & Services Administration, every day 17 people die waiting for an organ transplant in the US. This issue is now a public health crisis. Fortunately, due to the advancement of technology, three-dimensional (3D)-printed organs have become a reality.

In 2014, a California-based company called Organovo was the first to successfully engineer commercially available 3D-bioprinted human livers and kidneys. 3D printing in healthcare is used to create living human cells or tissues for regenerative medicine and tissue engineering purposes. The process of 3D printing typically begins with obtaining a sample of a patient’s own cells to grow and expand outside the body in a sterile incubator or bioreactor. These cells are then fed with nutrients called ‘media’ and mixed with a gel that acts as a glue. This mixture is then loaded into a printing chamber to build tissues by building the material up layer by layer.

Currently, the biggest challenge is to get the organs to function as they should. Despite the tremendous amount of progress being made in this field, Dr Anthony Atala and his colleagues at the Wake Forest Institute for Regenerative Medicine are conservative with their estimate about the number of years remaining before fully functioning 3D-printed organs can be implanted into humans.

In spite of the unknown timeline of when bioprinting organs can become an available option to patients, researchers are optimistic about the affordability of it for patients and their caregivers. The cost associated with organ failure is very high: just to keep a patient on dialysis is estimated to cost around C$350,000 ($270,000) in Canada, according to Ferguson and colleagues. According to research published by the American Society of Nephrology, in 2020 the average cost of a kidney transplant was $442,500, while 3D printers retail for upwards of $100,000, depending on their complexity. Adding costs of surgery and maintaining the 3D-printed organs could still be cheaper than a kidney transplant, according to Jennifer Lewis, a professor at Harvard University’s Wyss Institute for Biologically Inspired Engineering.

This is an exciting field that is still being developed and its speculated affordability is a good sign for patients and their caregivers.

Abstract

Biological materials are self-assembled with near-atomic precision in living cells, whereas synthetic 3D structures generally lack such precision and controllability. Recently, DNA nanotechnology, especially DNA origami technology, has been useful in the bottom-up fabrication of well-defined nanostructures ranging from tens of nanometres to sub-micrometres. In this Primer, we summarize the methodologies of DNA origami technology, including origami design, synthesis, functionalization and characterization. We highlight applications of origami structures in nanofabrication, nanophotonics and nanoelectronics, catalysis, computation, molecular machines, bioimaging, drug delivery and biophysics. We identify challenges for the field, including size limits, stability issues and the scale of production, and discuss their possible solutions. We further provide an outlook on next-generation DNA origami techniques that will allow in vivo synthesis and multiscale manufacturing.

The human brain contains functionally and anatomically distinct networks for representing semantic information in each sensory modality, and a separate, distributed amodal conceptual network. In this study we examined the spatial organization of visual and amodal semantic functional maps. The pattern of semantic selectivity in these two distinct networks corresponds along the boundary of visual cortex: for visual categories represented posterior to the boundary, the same categories are represented linguistically on the anterior side. These results suggest that these two networks are smoothly joined to form one contiguous map.

A unique form of brain stimulation appears to boost people’s ability to remember new information—by mimicking the way our brains create memories.

The “memory prosthesis,” which involves inserting an electrode deep into the brain, also seems to work in people with memory disorders—and is even more effective in people who had poor memory to begin with, according to new research. In the future, more advanced versions of the memory prosthesis could help people with memory loss due to brain injuries or as a result of aging or degenerative diseases like Alzheimer’s, say the researchers behind the work.

“It’s a glimpse into the future of what we might be able to do to restore memory,” says Kim Shapiro, a neuroscientist at the University of Birmingham in the UK, who was not involved in the research.

It works by copying what happens in the hippocampus—a seahorse-shaped region deep in the brain that plays a crucial role in memory. The brain structure not only helps us form short-term memories but also appears to direct memories to other regions for long-term storage.

Summary

Memristive devices share remarkable similarities to biological synapses, dendrites, and neurons at both the physical mechanism level and unit functionality level, making the memristive approach to neuromorphic computing a promising technology for future artificial intelligence. However, these similarities do not directly transfer to the success of efficient computation without device and algorithm co-designs and optimizations. Contemporary deep learning algorithms demand the memristive artificial synapses to ideally possess analog weighting and linear weight-update behavior, requiring substantial device-level and circuit-level optimization. Such co-design and optimization have been the main focus of memristive neuromorphic engineering, which often abandons the “non-ideal” behaviors of memristive devices, although many of them resemble what have been observed in biological components. Novel brain-inspired algorithms are being proposed to utilize such behaviors as unique features to further enhance the efficiency and intelligence of neuromorphic computing, which calls for collaborations among electrical engineers, computing scientists, and neuroscientists.