System Bits: Aug. 15

Machine-learning system for smoother streaming
To combat the frustration of video buffering or pixelation, researchers at MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) have developed “Pensieve,” an artificial intelligence system that uses machine learning to pick different algorithms depending on network conditions thereby delivering a higher-quality streaming experience with less rebuffering than existing systems.

“Studies show that users abandon video sessions if the quality is too low, leading to major losses in ad revenue for content providers. Sites constantly have to be looking for new ways to innovate,” according to MIT Professor Mohammad Alizadeh, whose team created Pensieve.

Sites like YouTube use adaptive bitrate (ABR) algorithms to try to give users a more consistent viewing experience. At the same time, it saves bandwidth: People usually don’t watch videos all the way through, and so, with literally 1 billion hours of video streamed every day, it would be a big waste of resources to buffer thousands of long videos for all users at all times.

In experiments, Pensieve could stream video with 10 to 30 percent less rebuffering than other approaches, and at levels that users rated 10 to 25 percent higher on key “quality of experience” metrics.
(Source: MIT CSAIL)

The researchers pointed out that while ABR algorithms have generally gotten the job done, viewer expectations for streaming video keep inflating, and often aren’t met when sites like Netflix and YouTube have to make imperfect trade-offs between things like the quality of the video versus how often it has to rebuffer.

The Pensieve AI system was found to be able to stream video with 10 to 30 percent less rebuffering than other approaches, and at levels that users rated 10 to 25 percent higher on key “quality of experience” (QoE) metrics.

In experiments, Pensieve could stream video with 10 to 30 percent less rebuffering than other approaches, and at levels that users rated 10 to 25 percent higher on key “quality of experience” metrics.
(Source: MIT CSAIL)

Pensieve’s neural network surveys the conditions of the user’s network in order to determine the appropriate bitrate for the situation. (Source: MIT)

Pensieve can also be customized based on a content provider’s priorities. For example, if a user on a subway is about to enter a dead zone, YouTube could turn down the bitrate so that it can load enough of the video that it won’t have to rebuffer during the loss of network, the team said.

Mimicking human thought
According to Purdue University researchers, a new computing technology called “organismoids” mimics some aspects of human thought by learning how to forget unimportant memories while retaining more vital ones.

Purdue postdoctoral research associate Fan Zuo, at left, and materials engineering professor Shriram Ramanathan, used a ceramic “quantum material” to create the technology. (Source: Purdue University)

Kaushik Roy, Purdue University’s Edward G. Tiedemann Jr. Distinguished Professor of Electrical and Computer Engineering explained, “The human brain is capable of continuous lifelong learning, and it does this partially by forgetting some information that is not critical. I learn slowly, but I keep forgetting other things along the way, so there is a graceful degradation in my accuracy of detecting things that are old. What we are trying to do is mimic that behavior of the brain to a certain extent, to create computers that not only learn new information but that also learn what to forget.”

Central to the research is a ceramic “quantum material” called samarium nickelate, which was used to create devices called organismoids. The work was performed by researchers at Purdue, Rutgers University, MIT, Brookhaven National Laboratory and Argonne National Laboratory.

“These devices possess certain characteristics of living beings and enable us to advance new learning algorithms that mimic some aspects of the human brain,” Roy said. “The results have far reaching implications for the fields of quantum materials as well as brain-inspired computing.”

When exposed to hydrogen gas, the material undergoes a massive resistance change, as its crystal lattice is “doped” by hydrogen atoms. The material is said to breathe, expanding when hydrogen is added and contracting when the hydrogen is removed.

The main thing about the material is that when this breathes in hydrogen there is a spectacular quantum mechanical effect that allows the resistance to change by orders of magnitude. This is very unusual, and the effect is reversible because this dopant can be weakly attached to the lattice, so if you remove the hydrogen from the environment you can change the electrical resistance.

Organismoids might have applications in the emerging field of spintronics. Conventional computers use the presence and absence of an electric charge to represent ones and zeroes in a binary code needed to carry out computations. Spintronics, however, uses the “spin state” of electrons to represent ones and zeros, the team said. This could bring circuits that resemble biological neurons and synapses in a compact design not possible with CMOS circuits. Whereas it would take many CMOS devices to mimic a neuron or synapse, it might take only a single spintronic device. In future work, the researchers said they may demonstrate how to achieve habituation in an integrated circuit instead of exposing the material to hydrogen gas.

RNA nanodevices in living cells
Synthetic biologists at the Wyss Institute at Harvard University are converting microbial cells into living devices that are able to perform useful tasks ranging from the production of drugs, fine chemicals and biofuels to detecting disease-causing agents and releasing therapeutic molecules inside the body.

To accomplish this, they said they fit cells with artificial molecular machinery that can sense stimuli such as toxins in the environment, metabolite levels or inflammatory signals. Much like electronic circuits, these synthetic biological circuits can process information and make logic-guided decisions. Unlike their electronic counterparts, however, biological circuits must be fabricated from the molecular components that cells can produce, and they must operate in the crowded and ever-changing environment within each cell.

So far, synthetic biological circuits can only sense a handful of signals, giving them an incomplete picture of conditions in the host cell. They are also built out of several moving parts in the form of different types of molecules, such as DNAs, RNAs, and proteins, that must find, bind and work together to sense and process signals. Identifying molecules that cooperate well with one another is difficult and makes development of new biological circuits a time-consuming and often unpredictable process.

The team at Wyss is now presenting an all-in-one solution that imbues a molecule of ‘ribo’ nucleic acid or RNA with the capacity to sense multiple signals and make logical decisions to control protein production with high precision.

The study’s approach resulted in a genetically encodable RNA nano-device that can perform an unprecedented 12-input logic operation to accurately regulate the expression of a fluorescent reporter protein in E. coli bacteria only when encountering a complex, user-prescribed profile of intra-cellular stimuli. Such programmable nano-devices may allow researchers to construct more sophisticated synthetic biological circuits, enabling them to analyze complex cellular environments efficiently and to respond accurately.

The team’s approach evolved from its previous development of so-called ‘Toehold Switches’ — first published in 2014 — which are programmable hairpin-like nano-structures made of RNA. In principle, RNA Toehold Switches can control the production of a specific protein: when a desired complementary ‘trigger’ RNA, which can be part of the cell’s natural RNA repertoire, is present and binds to the toehold switch, the hairpin structure breaks open. Only then will the cell’s ribosomes get access to the RNA and produce the desired protein.

“We wanted to take full advantage of the programmability of Toehold Switches and find a smart way to use them to expand the decision-making capabilities of living cells. Now with Ribocomputing Devices, we can couple protein production to specific combinations of many different input RNAs and only activate production when conditions allow it,” said co-first and co-corresponding author Alexander Green, Ph.D.
Green developed Toehold Switches with Yin and began the present study as a Postdoctoral Fellow in Yin’s team.

Illustration of an RNA-based ‘ribocomputing’ device that makes logic-based decisions in living cells. The long gate RNA (blue) detects the binding of an input RNA (red). The ribosome (purple/mauve) reads the gate RNA to produce an output protein. (Source: Alexander Green / Arizona State University)

Green is now Assistant Professor at the Biodesign Institute and the School of Molecular Sciences at Arizona State University where he continued experiments with his graduate student and co-author Duo Ma.