Why → Bio ML has been one of the most prominent areas of application and progress of machine learning techniques. This is a recent example of how generative models can be used in drug design based on protein bindings.
💡Key insights → The idea of Masked Language Modeling (corrupting a sequence by masking some elements and trying to reconstruct it) proves to be also useful to generate molecules: masking single atoms and ‘reconstructing’ how they should be filled up. By doing this process iteratively (e.g. ‘autoregressively’) this becomes a generative model that can output candidate molecules.
The main advancement of this paper with respect to previous works is that the generation of molecules is conditioned to particular protein binding sites, giving researchers a better mechanism to find molecule candidates to act as a drug for a particular purpose.
Source: https://papers.nips.cc/paper/2021/file/314450613369e0ee72d0da7f6fee773c-Paper.pdf
For this, several tricks are required for training, such as building an encoder that’s invariant to rigid transformations (translation and rotation), cause molecules don’t care about how they’re oriented. The results are not groundbreaking and there are still caveats to the method, such as the fact that it can often output molecules that are not physically possible, but this research direction is a certainly promising one for the future of drug development
Source: https://papers.nips.cc/paper/2021/file/314450613369e0ee72d0da7f6fee773c-Paper.pdf
Other works on ML for molecule structures at NeurIPS: Multi-Scale Representation Learning on Proteins, Hit and Lead Discovery with Explorative RL and Fragment-based Molecule Generation, Flow Network based Generative Models for Non-Iterative Diverse Candidate Generation.
2. The Emergence of Objectness: Learning Zero-shot Segmentation from Videos | 👾 Code
By Runtao Liu, Zhirong Wu, Stella Yu, and Stephen Lin.
Authors’ TL;DR → We present an applicable zero-shot model for object segmentation by learning from unlabeled videos.
❓Why → Humans can easily track objects which we have never seen and don’t recognize… Machines should do the same!
💡Key insights → The notion of objectness is often referred to as one of the important human priors that enable us to see and reason about what we see. The key here is that we don’t need to know what the object is to know it’s an object. On the contrary, when training a neural network for image segmentation in a supervised fashion, the model will only learn to segment objects seen during training.
This paper proposes to use video data and clever tricks to set up a self-supervised training approach that leverages how objects and foreground tend to behave differently in recordings. This is reminiscent of Facebook’s DINO [1], which uncovered how after training self-supervisedly with images, Transformers’ attention matrices resembled some sort of proto-segmentation.
In this case, the authors use a combination of single frame reconstruction loss (the segmentation network) and a motion network that tries to output a feature map of how pixels move in an image. This lets them predict a future frame given a reconstruction and a motion map, which is self-supervised end-to-end.
Source: https://papers.nips.cc/paper/2021/file/6d9cb7de5e8ac30bd5e8734bc96a35c1-Paper.pdf
Many more tricks and considerations are required to make this work, and the results are still not earth-shattering. Still, leveraging non-labeled data for Computer Vision using videos is a key development for the future of Machine Learning, so this line of work is bound to be impactful.
Other interesting Computer Vision papers at NeurIPS: Scaling Vision with Sparse Mixture of Experts (we already covered a few months ago!) Pay Attention to MLPs (also covered back in June), Intriguing Properties of Vision Transformers, Compressed Video Contrastive Learning,
3. Multimodal Few-Shot Learning with Frozen Language Models | 🌐 Website
By Maria Tsimpoukelli, Jacob Menick, Serkan Cabi, S. M. Ali Eslami, Oriol Vinyals, Felix Hill.
Authors’ TL;DR → We present a simple approach for transferring abilities of a frozen language model to a multi-modal setting (vision and language).
❓Why → Prompting is here to stay: now multimodal and async. Can you leverage the information in a pre-trained Language Model for vision tasks without re-training it? Well sort of… keep reading.
💡Key insights → The idea this paper proposes is fairly simple: train a Language Model, freeze it such that its parameters remain fixed, and then train an image encoder to encode an image into a prompt for that language model to perform a specific task. I like to conceptualize it as “learning an image-conditional prompt (image through a NN) for the model to perform a task”.
Source: https://papers.nips.cc/paper/2021/file/01b7575c38dac42f3cfb7d500438b875-Paper.pdf
It’s a promising research direction but the results are not very impressive (yet?) in terms of absolute performance. However, it is interesting to compare the model which is fully finetuned with multimodal data (Frozen finetuned) versus one that keeps the Language Model frozen (Frozen VQA-blind): only the latter shows good generalization from the training dataset (Conceptual Captions [4]) onto the target evaluation dataset (VQAv2 [5]), still being far from a fully supervised model.
Source: https://papers.nips.cc/paper/2021/file/01b7575c38dac42f3cfb7d500438b875-Paper.pdf
4. Efficient Training of Retrieval Models using Negative Cache
By Erik Lindgren, Sashank Reddi, Ruiqi Guo, and Sanjiv Kumar.
Authors’ TL;DR → We develop a streaming negative cache for efficient training of retrieval models.
❓Why → Negative sampling in dense retrieval is one of the most salient topics in the domain of neural IR! This seems to be an important step.
💡Key insights → A dense retrieval model aims to encode passages and queries into vectors to perform nearest neighbor search to find relevant matches. For training, a contrastive loss is often used where the similarity of positive pairs is maximized and that of negative pairs is minimized. Ideally, one could use a whole document collection as ‘negative samples’, but this would be prohibitively expensive, so instead, negative sampling techniques are often used.
One of the challenges of negative sampling techniques is that the final model quality strongly depends on how many negative samples are used — the more the better — but using many negative samples is computationally very expensive. This has led to popular proposals such as ANCE [2] where negative samples are carefully mixed with ‘hard negatives’ from an asynchronously updated index to improve a model performance at a reasonable computational cost.
In this work, the authors propose an elegant solution consisting of caching embeddings of documents when they’re calculated and only update them progressively, so instead of needing to perform a full forward pass in the encoder for all the negative samples, a large portion of cached documents are used instead, which is much faster.
This results in an arguably more elegant and simple approach that is superior empirically to existing methods.
Source: https://papers.nips.cc/paper/2021/file/2175f8c5cd9604f6b1e576b252d4c86e-Paper.pdf
Other information retrieval papers you might like at NeurIPS: End-to-End Training of Multi-Document Reader and Retriever for Open-Domain Question Answering, One Question Answering Model for Many Languages with Cross-lingual Dense Passage Retrieval, SPANN: Highly-efficient Billion-scale Approximate Nearest Neighborhood Search.
5. VATT: Transformers for Multimodal Self-Supervised Learning from Raw Video, Audio and Text
By Hassan Akbari, Liangzhe Yuan, Rui Qian, Wei-Hong Chuang, Shih-Fu Chang, Yin Cui, and Boqing Gong.
Authors’ TL;DR → A pure Transformer-based pipeline for learning semantic representations from raw video, audio, and text without supervision.
❓Why → Another step towards the promised multimodal future land.
💡Key insights → A still under-explored area for large transformer models to explore is that of multimodality. This work aims at constructing representations for video, audio, and text via self-supervised learning of data in those modalities using variants of contrastive losses jointly, so the three modalities inhabit the same embedding space. To do so, they use Noise Contrastive Estimation (NCE) and use matching audio/video-frame/text triplets as positive pairs and non-matching triplets (e.g. non-corresponding segments of a video) as negative samples.
This method is similar to previous multimodal networks such as “Self-Supervised MultiModal Versatile Networks” [3], with the main difference being that it relies on a pure Transformer architecture. It achieves SOTA in major video benchmark kinetics while avoiding supervised pre-training (e.g. using large scale Imagenet labels to pre-train the model), and using only self-supervised techniques.
Source: https://papers.nips.cc/paper/2021/file/cb3213ada48302953cb0f166464ab356-Paper.pdf
6. Robust Predictable Control | 🌐 Website
By Ben Eysenbach, Russ R. Salakhutdinov and Sergey Levine.
Authors’ TL;DR → We propose a method for learning robust and predictable policies in RL using ideas from compression.
❓Why → Excellent bird-eye paper to understand some of the fundamental challenges of Reinforcement Learning through the lens of compression.
💡Key insights → Despite environment observations being often high dimensional (e.g. millions of pixels from images), the amount of bits of information required for an agent to make decisions is often low (e.g. doing lane keeping of a car with camera sensing). Following this insight, the authors propose Robust Predictable Control (RPC), a generic approach to learning policies that use few bits of information.
To do so, they ground their analysis on the information-theoretical perspective of how much information a model needs to predict a future state: the more predictable a state the more easily a policy can be compressed. This becomes a regularisation factor for agents to learn to ‘play it safe’ most of the time (e.g. a self-driving system will tend to avoid high uncertainty situations).
Source: https://papers.nips.cc/paper/2021/file/e9f85782949743dcc42079e629332b5f-Paper.pdf
The specifics of the algorithm are much more complex so you’ll need to read the paper carefully if you want to dig into them. Beyond the particular implementation and the good empirical results, this seems to be a useful perspective to think of the tradeoffs one finds when navigating environments with uncertainty.
More on Reinforcement Learning at NeurIPS: Behavior From the Void: Unsupervised Active Pre-Training, Emergent Discrete Communication in Semantic Spaces, Near-Optimal Offline Reinforcement Learning via Double Variance Reduction.
7. FLEX: Unifying Evaluation for Few-Shot NLP | 📈 Leaderboard | 👾 Code
By Jonathan Bragg, Arman Cohan, Kyle Lo, and Iz Beltagy.
Authors’ TL;DR → FLEX Principles, benchmark, and leaderboard unifying best practices for evaluating few-shot NLP; and UniFew, a simple and strong prompt-based model by unifying pre-training and downstream task formats.
❓Why → Few-shot learning has been the NLP cool kid for some time now. It’s time it gets its own benchmark!
💡Key insights → Since GPT-3 [4] broke into the scene in May 2020 with surprising zero and few-shot performance on NLP tasks thanks to ‘prompting’, it has been standard for new large language models to benchmark themselves in the zero/few-shot setting. Often, the same tasks and datasets were used for finetuned models, but in this benchmark, zero/few-shot setting is a first-class citizen.
This benchmark coming from the Allen Institute for AI seeks to standardize the patterns seen in the NLP few-shot literature, such as The principles upon which this benchmark is built are: diversity of transfer types, variable number of shots and classes, unbalanced training sets, textual labels, no extra meta-testing data, principled sample size design and proper reporting of confidence intervals, standard deviations, and individual results. Hopefully, this will facilitate apples-to-apples comparison between large language models, which isn’t guaranteed when evaluation practices are a bit all over the place: the devil is in the details!
The authors also open-source the Python toolkit used to create the benchmark, along with their own baseline for this benchmark called UniFew, and compare it to a popular recent approach “Making pre-trained language models better few-shot learners” [5] and “Few-shot Text Classification with Distributional Signatures” [6].
Source: https://papers.nips.cc/paper/2021/file/8493eeaccb772c0878f99d60a0bd2bb3-Paper.pdf
Other NLP papers you might like at NeurIPS: COCO-LM: Correcting and Contrasting Text Sequences for Language Model Pretraining
8. Partition and Code: learning how to compress graphs
Authors’ TL;DR → We introduce a flexible, end-to-end machine learning framework for lossless graph compression based on graph partitioning, dictionary learning and entropy coding
❓Why → Compressing regular data (i.e. a sequence of symbols like 1s and 0s) has been extensively studied since Shannon’s introduction of Information Theory in 1948, however, compressing graphs has different peculiarities that are less well understood. Here’s a study on how to compress graphs from first principles.
💡Key insights → 3 peculiarities need to be considered when designing a compression algorithm for graphs:
Graph Isomorphisms: graphs don’t have an inherent ordering of vertices, unlike data sequences or arrays, so an optimally compressed codeword representation of a graph should be invariant to such isomorphisms.
Evaluating the likelihood of a graph: a theoretically optimal encoder relies on knowing the likelihood of possible each possible data configuration and assigning a code whose scale is proportional to the logarithm of such likelihood (e.g. more likely graphs get more compressed into shorter codes and vice-versa). Calculating such likelihood is generally intractable due to the combinatoric explosion of medium to large graphs! The problem needs decomposition…
If one builds a complex model to estimate the likelihood of graphs and compress based on that, the size of this model itself needs to be accounted for! Generally speaking, the more complex the model the better it’ll be able to compress data, but the larger it will be, thus incurring a tradeoff between how many bits one spends to store a model vs. how many bits one spends in storing each compressed graph instance.
While the authors don’t claim to propose an optimal solution, they present a practical one that loosely works as follows: partition graphs into common subgraphs for which you keep a dictionary of codewords whose length is proportional to the logarithm of each subgraph likelihood. The method is fully differentiable so it can be optimized with gradient descent for any given dataset of graphs.
Source: https://papers.nips.cc/paper/2021/file/9a4d6e8685bd057e4f68930bd7c8ecc0-Paper.pdf
When compared empirically to existing methods, it outperforms them, but it’s still an open question to what degree this will be used, given the introduced complexity. Regardless of their particular proposal, this is an excellent paper to understand how to compress things from the ground up.
Source: https://papers.nips.cc/paper/2021/file/9a4d6e8685bd057e4f68930bd7c8ecc0-Paper.pdf
Other interesting Graph Neural Network papers at NeurIPS: SIMONe: View-Invariant, Temporally-Abstracted Object Representations via Unsupervised Video Decomposition, VQ-GNN: A Universal Framework to Scale up Graph Neural Networks using Vector Quantization, GemNet: Universal Directional Graph Neural Networks for Molecules
9. Learning to Draw: Emergent Communication through Sketching | 👾 Code
By Daniela Mihai and Jonathon Hare.
Authors’ TL;DR → We use self-supervised play to train artificial agents to communicate by drawing and then show that with the appropriate inductive bias a human can successfully play the same games with the pretrained drawing agent.
❓Why → This one’s fun.
💡Key insights → Two models learn to communicate about images by drawing: a sender model needs to create a depiction of an image using a differentiable rasterizer that outputs ‘strokes’, and a receiver model needs to pick out which image the sender was representing out of a pool of images.
Source: https://papers.nips.cc/paper/2021/file/39d0a8908fbe6c18039ea8227f827023-Paper.pdf
An interesting observation is how without further constraints, the sender and receiver come up with drawing representations that are not human interpretable. But they try a clever regularization trick to incentivize human interpretability: adding a ‘perceptual loss’ at the early vision stage (i.e. in early layers of the encoder model) such that activations in the model for the original image and the drawing resemble each other. This is inspired by empirical observations of how neuronal activations in humans are similar for a given picture and a drawing of it.
Source: https://papers.nips.cc/paper/2021/file/39d0a8908fbe6c18039ea8227f827023-Paper.pdf
By Ruihan Wu, Chuan Guo, Awni Hannun and Laurens van der Maaten.
❓Why → The modules that form an AI system can interact in complex and unintuitive ways. This thought-provoking paper studies how making subparts of a system better can make the overall system worse.
💡Key insights → While this paper doesn’t propose any particularly impressive method, it’s interesting food for thought to always keep at the back of your head. The authors study and formalize the important problem of how and why a system composed of various ML subsystems might become worse when individual parts are improved. This is of high importance in the practical domain cause many AI systems are composed.
Source: https://papers.nips.cc/paper/2021/file/619427579e7b067421f6aa89d4a8990c-Paper.pdf
The theoretical analysis runs much deeper than this, but the main gist of it is that it’s probably true that you can degrade the performance of an AI system by improving each of its parts, so think twice before re-training that one ML pipeline component!
Here’s where our selection ends. Unfortunately, we couldn’t include many interesting works that were absolutely worthy of highlighting, so you’ll need to dive into the full list of conference papers to find them. For instance, you might want to check out the latest Schmidhuber’s Meta Learning Backpropagation And Improving It, or how Resnets are not dead Revisiting ResNets: Improved Training and Scaling Strategies, or the latest equivariance NN E(n) Equivariant Normalizing Flows? We’ll have to stop here out of respect for your time. We hope you enjoyed it, you can continue exploring the conference on our platform.
Download Bird mania
State Department Launches New Cybersecurity BureauThe bureau will bolster US diplomatic efforts to protect against cybersecurity threats from nations like Russia and China.April 4, 2022 by See full bioThe State Department on Monday launched a new cybersecurity bureau in an effort to make digital security a part of US foreign policy at a time when authoritarian regimes in Russia and China are increasingly trying to influence the internet.The move revives a bureaucratic structure that had been dismantled by the Trump administration, which was looking to cut red tape and streamline the diplomatic office. The new bureau will put in place a high-ranking cyber diplomat along with other diplomatic personnel and expertise to focus on addressing cyberthreats, global internet freedom, surveillance risks and to work with US allies to set international norms and standards on emerging technology. Secretary of State Antony Blinken had announced the establishment of the new Bureau of Cyberspace and Digital Policy at the State Department in October as part of an effort to reorganize and modernize the agency to meet 21st century needs. The new agency comes as Russia wages war in Ukraine and experts warn that the war could spur cyberwarfare meant to take out critical infrastructure like electrical grids, banks and communications. Such attacks can cause widespread destruction and are potentially deadly if they close hospitals and nursing homes. While no widespread attack has been waged thus far, the potential exists, experts say. Meanwhile, the US has long worried about cyberattacks from China. In July, the Biden administration blamed China for the cyberattack on the Microsoft Exchange email server software in early 2021. It said hackers linked to China’s Ministry of State Security exploited vulnerabilities to “compromise tens of thousands of computers and networks worldwide.”The US government has also adopted policies to limit or ban certain companies with ties to the Russian and Chinese governments from doing business in the US, for fear that the technologies from those companies could be used to spy on Americans or attack US communications networks. The new bureau is being established more than four years after then-Secretary of State Rex Tillerson merged the Office of the Coordinator for Cyber Issues with another State Department office. In the final days of the Trump administration, then-Secretary of State Mike Pompeo tried to reestablish the Bureau of Cybersecurity and Emerging Technology.The establishment of a new bureau within the State Department has gotten bipartisan support. Earlier this year, the House passed the Cyber Diplomacy Act, which would establish a cybersecurity bureau at the State Department and ensure it’s led by an official appointed by the president with the level of ambassador. The State Department said Monday that the new office will include three policy units: international cyberspace security, international information and communications policy, and digital freedom. Eventually, the bureau will be led by a Senate-confirmed ambassador-at-large. Until President Joe Biden appoints and the Senate confirms someone to this position, Jennifer Bachus, a career member of the Senior Foreign Service, will serve as principal deputy assistant secretary for the new bureau, starting Monday, the agency announced.
Cybersecurity warning ⚠
close
video
Russia-Ukraine: Cybersecurity expert warns US citizens to protect assets amid possible cyberattacks (video)
Cybersecurity expert and attorney Leeza Garber provides insight into how U.S. citizens can protect their assets and bank accounts during a possible cyberattack amid the Russia-Ukraine conflict.
The U.S. has seen a series of high-profile cyberattacks in recent years, from hackers hijacking systems for ransom payments to phishing scams targeting vulnerable businesses and individuals.
Yiddy Lemmer, who owns IT support and cybersecurity firm CompuConnect in New York City, says his company has seen a massive spike in attempted attacks against its clients in recent months and that the threats are “very, very real.”
The U.S. has seen a surge in cybersecurity threats against government entities, businesses and individuals. (iStock / iStock)
While he can’t say for certain who is behind all the breach attempts, Lemmer attributes the timing of the surge to Russia’s invasion of Ukraine and the U.S. sanctions against Russia that followed.
FBI DISRUPTS RUSSIAN MILITARY HACKERS, PREVENTING BOTNET AMID UKRAINE WAR
But regardless of cyber criminals’ intentions or locale, he says there are several ways businesses can bolster their cybersecurity to protect themselves.
Train employees on how to spot phishing scams
Lemmer says the vast majority of breaches occur through email, so being vigilant about threats and teaching employees about those dangers is the first step in being proactive against cyber threats.
Common scams used by malign actors include sending links through email that can hijack systems when clicked, posing as the worker’s boss to obtain banking information, or simply guessing easy passwords in order to gain access to company data.
Use strong, unique passwords and multi-factor authentication
Businesses and individuals alike should utilize password managers such as LastPass or Bitwarden that can generate secure passwords on the fly and store them for future use, Lemmer says.
He warns that using the same password for multiple sites leaves people vulnerable because of how frequently company sites are breached and their customers’ data end up for sale.
HACKERS TARGETED US ENERGY COMPANIES AHEAD OF UKRAINE INVASION: SOURCE
“Hackers will dip into that dark web page base to grab credentials and try to unlock as many doors as they can,” he told FOX Business.
Conduct a cybersecurity audit
Even companies with a dedicated IT specialist can have holes in their online protection, Lemmer says. He recommends having an independent third party conduct an audit to look for weaknesses, because firms like his that conduct such assessments inevitably find something that needs to be rectified.
Have proper backups in place
“One misconception people have is that keeping their files in the cloud means they do not need additional backups,” Lemmer told FOX Business. “The truth is that files in the cloud are just as susceptible to getting destroyed in cyberattacks and can get deleted remotely.”
HACKERS DUPED APPLE, META INTO HANDING OVER USER DATA: REPORT
He says there are direct-to-consumer services that provide secure online backup, and the one he recommends for consumers is Cloud Ally. But businesses should seek higher-grade protections that include compliance services.
( Jakub Porzycki/NurPhoto via Getty Images / Getty Images)
Obtain cyber insurance
Lemmer highly recommends that businesses obtain a cyber liability insurance policy, so that if they were to get hacked, the insurance company can help.
Insurance companies that offer these policies are incentivized to get involved not only to reduce or prevent the risk of any payouts for ransom demands but to assist in forensics to see how cybercriminals were able to gain access to a system in the first place.
However, Lemmer says firms that could benefit from those protections should act quickly. He explained, “More and more companies are stopping to offer cyber liability insurance because of the vast amount of cyberattacks happening.”
So, why is cybersecurity a big deal?
It is important to understand that cybersecurity includes both protecting digital forms of personal data and sensitive corporate information from exposure and protecting your business’ electronic systems from exploitation by hackers. The latter includes cases such as the Colonial Pipeline hack in 2021 when the company’s systems were taken over by Russia-based cybercriminals, ultimately resulting in Colonial Pipeline paying a $5 million ransom.
In addition to being exploited by cybercriminals, there are commonly legal repercussions with even minor cyberattacks. The most typical are investigations by regulatory agencies, breach notifications, and claims for damages for breach of contract and tort, but directors may also be held personally liable for cybersecurity breaches.
Can Directors be held personally liable for cybersecurity breaches?
Yes, a director can be personally liable for cybersecurity breaches in some instances. While no individual director has been held liable for a cybersecurity breach to date, lawsuits making these kinds of allegations have been filed, and it may be only a matter of time before one is successful. The primary risk of personal liability for a director is through derivative actions commenced by damaged shareholders. While the Business Judgment Rule generally insulates directors from personal liability, that protection is not absolute and can be rebutted.
Regulators and legislators are cracking down on cybersecurity practices
In addition to plaintiffs, regulators are also ramping up their response to cybersecurity breaches and increasing cybersecurity requirements for businesses. Numerous agencies have levied fines and brought suits for cybersecurity-related issues including the FTC, FCC, and SEC. For example, in 2014 the FCC fined two companies $10 million each for “unjust and unreasonable” data security practices in violation of the Communications Act of 1934.
In 2021, at least 45 states introduced or considered bills or resolutions concerning cybersecurity, more than 250 in total, and at least 35 states enacted cybersecurity-related laws. At the federal level, lawmakers have introduced at least 18 new bills concerning cybersecurity. For example, the Cybersecurity Disclosure Act of 2021 would require companies to disclose whether the board has any cybersecurity expertise or experience and, if not, report what aspects of the company’s cybersecurity were considered when evaluating nominees for membership on the board.\
What can a Board of Directors do to protect itself from cyberattacks?
Directors should prepare ahead of time to prevent the effects of cyberattacks and mitigate the risk of personal liability. Broadly speaking, boards must implement a reporting system and monitor or oversee the operation of that system to prevent personal liability under Caremark. In re Caremark Int’l Inc. Derivative Litig., 698 A.2d 959, 970 (Del. Ch. 1996). In Caremark, shareholders filed a derivative suit against the board after the company was required to pay approximately $250 million for violations of federal and state health care laws and regulations. Id. at 960–61. The Delaware Chancery Court held that directors can be held personally liable for failing to “appropriately monitor and supervise the enterprise.” Id. at 961. The court emphasized that the board must make a good faith effort to implement an adequate information and reporting system and that the failure to do so can constitute an “unconsidered failure of the board to act in circumstances in which due attention would, arguably, have prevented the loss.” Id. at 967. While Caremark did not address cybersecurity directly, the court’s reasoning in Caremark is applicable to board involvement, or lack thereof, with cybersecurity.
In addition, businesses and boards should implement greater protections to avoid further liability under other legal theories such as negligence. For example, as a result of the Colonial Pipeline hack, plaintiffs have filed a class action complaint asserting a negligence claim against the owners of the pipeline for failing to prevent the hack. Dickerson v. CDPQ Colonial Partners, L.P., Case No. 1:21-cv-02098-MHC, WL 2009109 (N.D. Ga. 2021).
10 Questions Boards Should Ask to Protect Against Cyberattacks:
What are our most important assets and what are our greatest cybersecurity risks?
What is our cybersecurity and data protection plan?
Does our cybersecurity system work across all of our platforms, devices, tablets, phones, and laptops including personal devices?
What is our response plan?
What is the board’s role in the event of a cybersecurity incident?
Is our cybersecurity investment enough?
10 Steps Boards Should Take for Protection:
Ensure there is a cybersecurity expert on the board (see pending House bill, Cybersecurity Disclosure Act of 2021);
Hire a Chief Information Officer (CIO) and/or a Chief Information Security Officer (CISO) – having the right people in place is crucial;
Engage outside experts to conduct regular cybersecurity assessments, including penetration testing;
Educate directors, officers, and all other employees about cybersecurity;
Regularly address and deliberate about cybersecurity issues and document discussions; having an incident response plan in place before a crisis means that the company is better able to respond and minimize the impact of a cyber incident;
Adopt and employ a cybersecurity system and plan that is specifically tailored to your business’s most important assets and risks;
Obtain cybersecurity insurance and ensure that it adequately covers potential risks for your company;
When a breach occurs, determine the scope of the breach, assess exposure, and comply with notification requirements.
Article written with assistance from Cranfill Sumner LLP clerk Devin Honbarger.
Amazon’s Project Kuiper is like SpaceX’s Starlink which provides satellite internet!
Sustainable cybersecurity starts with protecting both sides of the entry pointApril 5, 2022 by Ken Galvin2021 was a landmark year for cybersecurity — and not in a good way. Squaring up against broadening IT ecosystems, emerging threat vectors and innovative bad actors, security practitioners have been dealt a heavy hand in the past 24+ months, and if Log4j taught us anything, it’s not looking like threats will be easing up anytime soon. But as organizations look to bolster their cyber resiliency and maintain customer trust in the new year, it’s imperative that they consider cybersecurity today is not just about shoring up the endpoint; it’s just as vital to ensure your users are acting conscientiously. One false click, one false move on the user end, and your security can go out the window. To really be resilient, organizations need to understand that sustainable cybersecurity is really about protecting both sides of the entry point.Let’s look at data, for example. It’s now an enterprise given that “data intelligence” is vital to helping enterprises transform themselves to be more innovative, competitive and customer-centric. In other words, data is the lifeblood of business today. Not only does understanding data and data systems empower everyone in the organization to make faster, smarter and more strategic decisions to drive business growth, but it also plays a vital role in propelling innovation. So, protecting the gateways to your data — your endpoints — is more critical than ever before. And if you can’t manage your endpoint security — you can’t manage the security of your data, or have data privacy.For organizations looking to tighten up their cybersecurity this year, and ensure all entry points are protected and accounted for in 2022, here are a few key considerations to keep top of mind: Discovery. Ask yourself: What is hitting your network? If you do not know who or what is potentially accessing your data, you cannot secure it. You need to be able to use a variety of methods to not only discover and inventory devices that should be there, but also to detect devices that should not be.Vulnerability scanning. Not just for Windows, but also Mac and Linux systems. Most cyberattacks involve unpatched operating systems and applications that have had patches available for months — even years. Use your scan results to automatically isolate vulnerable devices and quickly remediate issues.Secure everything. Just as laptops pose their own unique access and security concerns, non-computer devices are also potential threat vectors. Ignoring the security of all your IoT devices (the fastest-growing device segment) is a disaster waiting to happen. Ensure that you’re constantly securing and assessing everything — on both sides of the endpoint — to achieve sustainable cybersecurity resiliency within your organization. In short, we can expect that the security trends that came with 2021 — upticks in insider threats, a rise in ransomware attacks, and unprecedented data breaches — are here to stay in 2022. In order for organizations to bolster their cybersecurity resiliency and prepare for the onslaught of attacks and cybersecurity threats to come, it’s imperative that they consider that endpoint security today is a two-way street. Sustainable security starts with shoring up both sides of the entry point
5 cybersecurity mistakes you’re making.
5 dangerous cybersecurity mistakes you’re probably making
NEWYou can now listen to Fox News articles!
The FBI’s latest Internet Cybercrime Report paints a bleak picture. Last year, Americans lost more than $6.9 billion to cybercrooks.
Don’t think you’re too smart to fall for their tricks. Even savvy people can get conned out of money before realizing what happened. Tap or click five simple, effective ways to secure your smartphone.
Maybe it’s too late, and you’ve noticed unexpected pop-ups or your phone getting hot when you’re not using it. Here’s how to know if a hacker or snoop is already in your smartphone.
Avoiding cybercriminals feels like a feat, but it doesn’t have to be that complicated. Knowledge is power. I’ll walk you through five mistakes you might be making.
1. You think free means safe
Taking advantage of “free” Wi-Fi may cost you more than money. Public networks are unsecured and easy to hack. I’m not just talking about airports. Your local coffee shop, salon, or any place that doesn’t password-protect its network leaves you and your data vulnerable.
Since this network is open for use, packet sniffers are readily available online that capture every keystroke you type. Think about this. Your passwords can be seen and collected by criminals.
Use a virtual private network when you need to access the internet and are away from a secure wireless network. A VPN uses an encrypted connection to safeguard against snoops.
You can also use your phone as a hotspot. Tap or click here for iPhone instructions. For steps on an Android, tap or click here.
2. You skip updates
Are you notorious for rescheduling software updates but never actually installing them? If you often hit the “Remind me later” button, you ask for trouble. Don’t prevent your system from receiving the latest tools and security patches needed to fight off attackers and malware.
Young girl is crying lonely with tears in front of her laptop late at night (iStock) (iStock)
Updates are annoying when you’re in the middle of your workday, so schedule them late at night when you’re not using your computer. Tap or click here to schedule updates on your Windows PC.
3. You pick up when a scammer calls
Sometimes those scam numbers are mighty convincing. You recognize the area code and maybe even the first few numbers, or perhaps it’s your phone number. You pick up. That’s when a scammer has a chance to get their claws into you.
If you see Scam Likely, or whatever term your carrier and phone display, don’t answer. I often hear from my national radio show listeners who like to play games with phone scammers. They egg them on and pretend they’re interested.
This isn’t too bright. You never know if that person is recording your voice for nefarious purposes or even making a deepfake audio recording of you later.
Like what you’re reading? Find the radio station near you that airs my show. You can also get the show’s podcast commercial-free.
The more online accounts you have, the more at risk you are when hackers come calling. With a new breach around every corner, your usernames and passwords aren’t safe.
Step one is combing through your email inbox and phone to locate the accounts you’re not using anymore. Then get rid of them. That’s not always the easiest thing to do.
FILE – The icons of Facebook and WhatsApp are seen on an iPhone in Gelsenkirchen, Germany, Nov. 15, 2018. (AP Photo/Martin Meissner, File)
Some accounts are impossible to delete, and some sites hide their delete links, and you have to dig pretty deep to find them. Tap or click for a tool that makes it easy to find exactly where to cancel online accounts.
This takes some time, but it’s worth it. When the inevitable data breach is announced from a site you once used, you’ll be glad you did it.
5. You click agree
When was the last time you read a site or service’s terms and conditions? You’re not alone. It probably means you’re allowing companies to collect your private data.
I’m not proposing you read every word because I know that’s not realistic. But there is a clever way to at least check a few things.
Next time you run into a privacy policy, terms and conditions page, or lengthy terms of service agreement, use a keyboard shortcut to search for specific words.
On a Windows PC, use Control + F.
On a Mac, use Command + F.
Now, type in terms like “third party,” “GPS,” “tracking,” and “data.” You’ll get a quick look at how your info is being used.
A woman in front of a computer. (iStock)
Bonus Tip: Wi-Fi on the moon, Russian cyberattacks and a Google tip for saving gas money
Did you know Wi-Fi is coming to the moon? Yes, really. In this episode of Kim Komando Explains, I’ll teach you how to find an airline seat with the most legroom, save gas money and a few other tech tips you’ll use time and time again. I’ve also got an action plan you can use to protect yourself against Russian cyberattacks.
Check out my podcast “Kim Komando Explains” onApple, Google Podcasts, Spotify, or your favorite podcast player.
Listen to the podcast here or wherever you get your podcasts. Just search for my last name, “Komando.”
What digital lifestyle questions do you have? Call Kim’s national radio show and tap or click here to find it on your local radio station. You can listen to or watch The Kim Komando Show on your phone, tablet, television, or computer. Or tap or click here for Kim’s free podcasts.
Copyright 2022, WestStar Multimedia Entertainment. All rights reserved. By clicking the shopping links, you’re supporting my research. As an Amazon Associate, I earn a small commission from qualifying purchases. I only recommend products I believe in.
Learn about all the latest technology on The Kim Komando Show, the nation’s largest weekend radio talk show. Kim takes calls and dispenses advice on today’s digital lifestyle, from smartphones and tablets to online privacy and data hacks. For her daily tips, free newsletters, and more, visit her website at Komando.com.
Artificial cybernet updates. info 